Controlled Release Strategies for Bone,Cartilage,and Osteochondral Engineering—Part II: Challenges on the Evolution from Single to Multiple Bioactive Factor Delivery

Single GF delivery

Administration of GF and other bioactive molecules to promote bone and cartilage formation and repair has achieved promising results. Several controlled release systems promoting the delivery of GFs are combined in strategies with cells, with the goal of acting in a synergistic manner and promoting the enhancement of new tissue formation. Co-encapsulation of GFs and cells in hydrogels^2–10 and seeding of stem cells in microparticles or scaffolds loaded with bioactive agents^11–18 are among the most common TE strategies described in literature. Typically, GFs have been included in TE strategies through three main approaches: (1) incorporation within micro- and nano-particles, which can act as supplements for in vitro cell cultures or can be injected into the defect sites, stimulating in situ tissue healing. These carriers can also be cell seeded, further enhancing the potential for inducing new tissue formation.^{11,15,19–31} (2) GF-loaded microspheres can be incorporated into scaffolds or hydrogels to enhance their functionality and complexity, providing the biochemical cues to stimulate tissue regeneration.^12,32–50 (3) Bioactive factors directly dispersed, immobilized, or adsorbed into the three-dimensional construct.^{3,4,6–8,51–68} The level of immobilization determines the release rate of GFs and consequently their effect on tissue formation. Several studies approach a particularly relevant immobilization method, mimicking native extracellular matrix (ECM), affinity-bound systems, through the inclusion of heparin domains on the structure, thus expecting an enhanced stability of the entrapped GF.^{27,66,69–72}

Table 1 presents selected studies regarding the use of single GF release systems for bone and cartilage engineering. Current evidence based on in vitro and animal studies suggests that among the factors that have been investigated to date, bone morphogenetic proteins (BMPs) appear to have the highest osteoinductive potential. ⁷³ Despite the successful reports related to the delivery of BMP for bone TE applications, when translating into clinical trials, the outcomes have not been as successful. Large doses of the GF have been required to produce an osteogenic effect.^73–79 The use of supraphysiological levels of BMP-2 might activate a negative feedback loop through BMP inhibitors such as noggin or sclerotin, which are upregulated by the presence of BMP.^80,81 Moreover, inflammation, edema, and heterotopic bone formation can occur when such high concentrations are used. ⁸² Complications associated to the use of recombinant human BMP-2 have been reported, including death, dysphagia, heterotopic bone formation in the spinal canal, or airway compression in cervical spine fusion. Food and Drug Administration (FDA) has been especially cautious since those reports and at this point, 75% of the spine fusions are still performed by using traditional bone grafting methods. ⁸¹ Therefore, despite its tremendous potential, BMP-2 use still presents some disadvantages regarding its bone regeneration potential. The high-cost treatment and the simultaneous stimulation of development of both osteoblasts and osteoclasts with opposite effects are additional drawbacks. ⁸³

Despite the complexity of angiogenic signaling pathways, extensive studies on the biology and delivery of vascular endothelial growth factor (VEGF) and fibroblast growth factor (FGF)-2, and potent GFs involved in vessel formation have been performed in both preclinical and small-scale clinical trials. Recent large-scale clinical trials ⁸⁴ did not demonstrate significant therapeutic effects above placebo controls, leading to an inefficient promotion of angiogenic response. A threshold concentration of VEGF is required, above which saturation of receptors occurs and consequent down regulation of receptor expression. On the other hand, low levels of VEGF lead to insufficient expression.

Appropriate in vitro or in vivo models allowing quantitative characterization of the effects of GFs in neovascularization are crucial for the development of appropriate design criteria for therapies. In vivo models, including chick chorioallantoic membrane assay, corneal assay, dorsal skin chamber assay, subcutaneous implantation, and induced ischemia models, mimic certain aspects of human diseases. ⁸⁴ In vitro models such as the embedding of endothelial cells in fibrin, ⁸⁵ collagen, ⁸⁶ or Matrigel ⁸⁷ hydrogels can also mimic some of the in vivo microenvironmental features occurring during angiogenesis.

The kinetics of release strongly influences tissue regeneration. Extremes of release are in general not desirable and controlled and prolonged profile should be designed. The work of Jeon et al. ⁷⁰ showed that prolonged delivery of BMP-2 enhanced in vivo the osteogenic efficacy of the protein compared to short-term delivery at equivalent dosage. Moreover, the physiological angiogenic response requires a precisely coordinated interplay between different signaling molecules and cell types. Ozawa et al. ⁸⁸ showed that the key determinant whether VEGF-induced angiogenesis is normal or aberrant is the microenvironmental amount of GF secreted, rather than the overall dose. Other approach used to promote angiogenesis is the blocking of angiogenesis inhibitors. ⁸⁹ Pro-angiogenic and anti-angiogenic factors exist in dynamic balance, and when angiogenesis inhibitors such as thrombospondin-2 are knocked out, increase in fracture vascularization can be observed. ⁹⁰

One of the critical issues in GF delivery has been the appropriate tailoring of signaling molecules dosage. There is almost no common ground in this matter and typically the use of GFs is observed within broad ranges of concentrations. For example, from our collected data in Table 2, BMP-2 and transforming growth factor (TGF)-β1 have been used in vivo for bone and cartilage regeneration within a range of 0.015–150 μg and 0.8 ng–1 μg per implant, respectively.

These differences cannot be explained only by animal models or discrepancies on the type of defects and ultimately emphasize the lack of knowledge regarding the therapeutic concentration of GFs. In fact, the FDA-approved BMP factors have been used in supraphysiological amounts to obtain a therapeutic effect and the required excessive amounts of proteins can be explained by an ineffective delivery system, which does not present the factor in a spatiotemporal controlled manner.

Tissue development is a highly coordinated process and it significantly benefits from strategies that mimic concomitant interactions among various factors involved in this process. ¹²² The presentation mode of GFs is a critical step for the generation of a specific tissue response, and several studies have focused on the evaluation of the effect of immobilization technique on cellular response.^123–131 For example, it is reported that soluble VEGF induces endothelial cell proliferation, while matrix-bound VEGF promotes vascular sprouting and branching ¹³² despite other studies revealing opposite responses.^{86,130,133,134} Moreover, sustained activation of the Smad intracellular signaling pathway is stimulated upon culture of osteoblastic cells on immobilized BMP-2 due to prolonged phosphorylation of Smad 1/5/8 by cellular exposure to immobilized BMP-2 relative to treatment with soluble BMP-2. ¹³⁵

Several studies^{127,136–138} have shown that the culture of osteoblast precursor cells on substrates modified with immobilized BMP-2 can significantly increase the expression of osteogenic differentiation markers. These immobilization techniques can even be combined with patterning approaches. ¹³⁹ In particular, immobilization with affinity-bound ligands has been demonstrated to promote a strong attachment of GFs to the matrix and to enhance their potency. For example, the osteoinductive effects of recombinant human BMP-2 in combination with a complex of heparin and chitosan in a gel formulation were shown to be superior when compared to recombinant human BMP-2 implanted with type I collagen in a rat model. ¹⁴⁰ Jeon et al. ⁷⁰ also observed that the heparin-conjugated scaffolds allowed a long-term delivery of BMP-2, which ultimately resulted in the enhancement of the in vivo osteogenic efficacy of the protein.

Re'em et al. ¹⁴¹ also showed that controlled release of TGF-β1 affinity-bound alginate scaffolds enhanced human mesenchymal stem cells (MSCs) chondrogenic differentiation and in vivo deposition of cartilaginous ECM in an ectopic model in nude mice. Reyes et al. ¹⁴² analyzed osteochondral regeneration postimplantation of a bilayered scaffold loaded with BMP-2 or TGF-β1 and observed that the higher concentration of BMP-2 gave rise to higher quality cartilage with improved surface regularity 2 weeks postimplantation.

Despite the partial success of simple GF therapy, it is clear that the field is not taking full advantage of the potency of these signaling molecules. The complexity of native tissue-healing cascades, which involves several GFs and chemokines, communicating with each other through positive and negative feedback mechanisms, is obviously lacking in current TE and regenerative strategies incorporating the release of a single GF. Even if the carrier and release kinetics are appropriately designed, a single signaling molecule will not be able to promote bone and cartilage regeneration by itself. Therefore, the development of TE strategies incorporating delivery systems for multiple GFs has emerged to overcome the hurdles previously found and to increase the functionality of the constructs following a biomimetic approach.

Multiple GF delivery

Appropriate delivery of multiple GFs and other bioactive agents might reduce the dosage of factors required to achieve the desired effect, in essence increasing the potency of the molecules. Precise control over temporal sequence of release and presentation of GFs is critical because coexistence of destabilizing and stabilizing factors may cancel each other's effects. ⁸⁴ The hallmark study of the use of dual and multiple GF release for TE was the one performed by Richardson and his team. ⁴⁴ Angiogenesis is one of the most relevant mechanisms involved in bone healing, which is characterized by complex cascades of GFs. VEGF in particular has shown its potency to promote therapeutic angiogenesis. However, the vessels induced by single delivery of VEGF frequently display morphological and functional abnormalities, such as leaky vessels and unusually large irregular lumens. ⁸⁸ Richardson et al. ⁴⁴ successfully designed a VEGF and platelet-derived growth factor (PDGF)–BB dual release system with distinct delivery profiles, promoting a rapid generation of mature vascular network. Moreover, they showed the thin line between the therapeutic effect of a successful combination of GFs, delivered at the appropriate time and dosage, and the antagonist effect of this combination. In this case, high levels of PDGF before sufficient pericyte recruitment result in destabilized vessel network and subsequent regression. Upon in vivo implantation, the mechanism of dual delivery allows the formation of larger and more mature blood vessels as opposed to smaller, incomplete vessels formed using a single deliver technique. ¹⁴³ Moreover, formation of truly functional vasculature will likely require control over the location and the magnitude of the angiogenic region. Tight spatial regulation often results from the combined action of stimulatory and inhibitory factors. ¹⁴⁴ Yuen et al. ¹⁴⁴ employed a dual-release system based on a poly(lactic-co-glycolic acid) (PLG) scaffold incorporating layers loaded with angiogenic stimulatory factor and inhibitory anti-VEGF. This led to a spatially sharp angiogenic region, sustained over 3 weeks.

Table 2 presents a detailed description of some of the most relevant studies performed regarding the application of dual or multiple bioactive factor delivery systems for bone, cartilage, and osteochondral regeneration.

The choice of the appropriate combinations of GFs is one of the critical hurdles for bioactive factor delivery in TE. ⁸² As an example, Vonau et al. ¹⁴⁵ designed a dual-delivery system composed by recombinant human BMP-2 and FGF-2 in a collagen sponge, which ultimately resulted in decreased bone formation in a rabbit model of tibial fracture.

The modulation of the adequate delivery kinetics is another major design requirement that can significantly affect the outcome of the strategy. Setting and combining fast and slow-release profiles can enhance tissue formation by closely mimicking native interactions in ECM. For example, Jaklenec et al. ¹⁴⁶ designed modular scaffolds resultant from the fusion of PLGA microparticles, tailored for different delivery rates of GFs (delayed and burst release), allowing sequential delivery of insulin growth factor (IGF)-I and TGF-β1.

On a first analysis of Table 2, the diversity of combinations and dosages of GFs is clear. Furthermore, the use of different materials processed in distinct architectures and the specific drug-loading procedures do not contribute for the homogenization of the outcomes. One obvious limitation of several studies that assessed their drug delivery system in vivo is the in vitro quantification of the release profile of the bioactive factors and consequent assumption that the kinetics would follow a similar pattern in vivo. These observations rise obstacles for the establishment of a correlation between the delivery kinetics and the measured outcomes and decreases significantly the efficiency of predictability of a newly proposed drug delivery system.

Regarding the selection of bioactive factors, since the lack of bone formation is often due to the limited ability of the surrounding tissue to induce a vascular supply at the regenerating location, one of the most common combinations for bone regeneration is the dual release of osteogenic and angiogenic GFs, since those two processes are interconnected during bone healing. Vessel formation is the earliest process in bone regeneration to promote the recruitment of progenitor cells to start the osteogenic differentiation process. ¹⁷⁷

Dual immobilization has also been performed to further improve angiogenesis. VEGF has been coimmobilized with angiopoietin-1 (Ang-1), a GF known for its support for vessel stability and maturation, and enhanced endothelial cell infiltration was achieved. ¹²⁶

The combined release of BMP-2 and VEGF is commonly approached with mixed results, and control over amount and timing of GF delivery is critical. VEGF dosage must be tightly controlled as excessive amounts of this GF can inhibit osteogenesis and cause severe leakage and hypotension. ¹⁶⁷ Combined results of Kempen et al. ¹⁶⁵ and Patel et al. ⁴⁵ suggest that the enhanced effect of VEGF and BMP-2 combination is both time- and location-dependent, which is not surprising due to the complexity of the pathways involved in bone healing. While Kempen et al. ¹⁶⁵ displayed an extremely high VEGF burst release of around 80%, De la Riva et al. ¹⁶⁶ reduced this stage to 20% and obtained considerable enhanced bone formation in the dual-release system they proposed (combination of VEGF and PDGF). Shah et al. ¹⁶⁷ showed that dual release of VEGF and BMP-2 from polyelectrolyte multilayers induced a more complete bone architecture than the single dose of BMP-2, promoting a greater initial concurrent vascularization process and consequent introduction of more cells in the interior on the scaffolds. However, a recent report Geuze et al. ¹⁷⁰ demonstrated that the timing of BMP-2 release largely determined the rate and amount of bone formation independently of VEGF release kinetics. Moreover, at orthotopic location, no significant effect on bone formation was found from a timed release of GFs, suggesting that time-release effects are location dependent.

Other combinations have also been explored to promote bone regeneration. Luong et al. ¹⁷⁸ developed an interesting study on the effects of the coprecipitation of different amounts and combinations of FGF-2 and BMP-2 into an apatite coating on the osteogenic differentiation of bone marrow mesenchymal stem cells (BMSCs). These studies revealed that FGF-2 increased cell proliferation, but high concentrations inhibited osteogenic differentiation. On the other hand, a threshold concentration of BMP-2 was required to induce significant osteogenesis. The sequence of delivery of BMP-2 (300 ng/mL) and FGF-2 (2.5 ng/mL) also provided relevant data, and it was observed that the release of FGF-2 followed by BMP-2 or the delivery of BMP-2 followed by the simultaneous delivery of FGF-2 and BMP-2 enhanced osteogenic differentiation more significantly than the simultaneous delivery of both factors. ¹⁷⁸

Strobel et al. ¹⁷⁹ applied a multiple and sequential delivery coating, promoting the release of gentamicin, IGF-1, and BMP-2. The earlier action of IGF-1 promoted cell proliferation, while BMP-2 stimulated alkaline phosphatase (ALP) activity. Kohara and Tabata ¹⁶⁸ also explored an interesting route by developing a dual-release system based on gelatin hydrogels loaded with calcium phosphates and with BMP-2- and Wnt1-inducible signaling pathway protein 1. The Wnt-signaling pathway is predicted to control bone mass, and the dual-release system was evaluated in middle-aged mice with decreased bone formation potential. The simultaneous delivery of both factors increased osteoid formation.

One interesting finding of Ripamonti et al. ⁶⁰ was the apparent role of osteoclast activity on the implanted macroporous constructs. In their study, bisphosphonate zoledronate was loaded onto the surface of the scaffolds to inhibit osteoclastic activity after intramuscular implantation. Osteoclasts surfacing and modifying the surface of the implanted constructs might initiate bone formation by carving topographical modifications and releasing calcium ions. The inhibited scaffolds showed lack of new bone formation, indicating the critical role of osteoclasts in the spontaneous induction of new bone formation.

Core-shell microcapsules using PLGA and alginate were also developed for dual bioactive factor-release system. The core and shell domains presented different release patterns and when incorporating BMP-2 and dexamethasone and co-cultured with BMSCs in a hydrogel construct resulted in enhanced expression of osteogenic markers. ¹⁸⁰ It has also been reported that the coadministration of VEGF, IGF-1, stromal-derived growth factor (SDF)-1, and Ang-1 led to a dramatic increase in angiogenic response, specifically the size and number of arterioles when compared to the single delivery of these individual factors. ¹⁵⁴

Another interesting group of studies involve the use of chemoattractant chemokines selected to promote host cell migration. For example, SDF-1 chemotactic gradients have been shown to affect migration patterns of both injected and host MSCs.^181–183 Therefore, the targeted delivery of SDF-1 can be used as a new approach to create an artificial signaling center without implantation of exogenous MSCs. Despite the potential of inflammation modulation, the critical situation is to distinguish between regenerative and damaging inflammation processes in bone. Aberrant inflammation has been implicated as a significant factor in bone injuries that fail to heal; thus, tight control over this process must be tailored. ¹⁸⁴

An obvious interesting approach for a multiple delivery system is the regeneration of orthopedic interfaces, such as ligament-to-bone,^185,186 tendon-to-bone, ¹⁸⁷ and cartilage-to-bone. ¹⁸⁸ Many approaches have been used to fabricate scaffolds for osteochondral application by changing material composition, mechanical properties, and architecture between the chondrogenic and osteogenic layers.^{172–174,188–193} Hierarchical scaffolds loaded with inductive factors disposed in two phases or even in gradients, capable of stimulating each layer toward the maturation of the specific tissue, are being increasingly explored. By using gradients of multiple bioactive factors, multiple tissue regeneration can be addressed via a single-cell source; that is, stem cells can differentiate along different lineages within the same constructs. ¹⁹⁴ Current design challenges for engineering biomimetic gradients are caused mostly by a question of scale, as it is not an easy task to mimic the micro- and nanoscale gradients reported at tissue-to-tissue interfaces, such as the osteochondral one. ¹⁹⁵

The application of dual-delivery systems for cartilage regeneration has also been under intensive investigation. Bian et al. ³⁶ successfully implanted subcutaneously in nude mouse constructs containing TGF-β3-loaded microspheres, resulting in superior cartilage matrix formation when compared to groups without the GF or with the protein added directly to the gel. However, calcification was observed, hence suggesting hypertrophy of chondrocytes. To counter this aberrant hypertrophy, parathyroid hormone-related protein (PTHrP) has been employed to inhibit hypertrophy of chondrocytes or MSCs during chondrogenesis. To prevent this, the authors implanted a dual-delivery system releasing both TGF-β3 and PTHrP, resulting only in partially reduced calcification and failing to prevent extensive mineralization. These results might be explained by the quick release of the hormone during the first week of implantation, and the system should be optimized to delay the release of PTHrP until the second week of implantation, when MSCs begin to express hypertrophic markers.

The combination of TGF-β1 and IGF-1 has been intensively explored by Mikos and coworkers.^171,172 Codelivery of both GFs resulted in enhanced expression of chondrogenic markers in some of the studies; however, in vivo data also showed that the dual-delivery system did not maintain the positive effect of IGF-1 as a single-bioactive-factor delivery system. These observations emphasize that even if the sequence of GFs may be appropriate, dosage and release kinetic optimization are required.

The application of dual-release systems for osteochondral regeneration has also been increasingly reported. One of the most interesting findings of the study of Holland et al. ¹⁷² was the apparent lack of TGF-β1 and IGF-1 synergy when simultaneously delivered to treat rabbit osteochondral defects. The GFs were delivered at different rates, with IGF-1 released throughout the first weeks of healing, while TGF-β1 was expected to be released within the first few days. The individual release of IGF-1 consistently produced better regeneration outcomes, therefore showing the need to optimize the proper combination, dosage, and release kinetics of different GFs to achieve the desired biological effect. Mohan et al. ¹⁷⁴ also observed that sintering BMP-2- and TGF-β1-loaded microspheres disposed in oppositely oriented gradients could lead to appropriate scaffolds aiming for osteochondral regeneration, as demonstrated in vivo through implantation in rabbit condyle defects.

Wang et al. ¹⁷³ investigated the release of BMP-2 and IGF-I inside PLGA and silk fibroin microspheres, suspended in a gradient pattern inside alginate gels and observed a direct effect over osteogenic and chondrogenic differentiation of seeded hMSCs. Dormer et al. ¹⁷⁵ designed PLGA microsphere-sintered constructs incorporated with gradients of BMP-2 and TGF-β1, and despite obtaining promising data, acknowledged the difficulty in sustaining the gradient patterns for long term in vitro due to the quick diffusion of the molecules. Re'em et al. ¹⁷⁶ designed a bilayered system, spatially presenting the chondroinductive TGF-β1 in one layer and the osteoinductive BMP-4 in a second layer via affinity binding to the matrix. When implanted in subchondral defects in rabbits, the constructs were able to induce the migration of host stem cells that sensed the biological cues spatially presented in the hydrogel and responded by differentiating into the appropriate cell lineage.

Another rather important conclusion taken from the compilation of studies focusing on the engineering of drug delivery systems for bone, cartilage, and osteochondral applications is the lack of in vivo proper tracing and quantification of the bioactive factor being released in the host. Most of the reports that assessed the in vivo release profiles used radioactivity as a tracing agent to identify the implanted molecules of interest.^160,166,168

Indirect GF delivery

The development of systems capable of promoting the delivery of inductive molecules has seen a tremendous improvement over the last decades. A key challenge in the application of GFs is their eventual inefficient delivery. In this review, we classify indirect growth delivery strategies as the ones leading to the production of these bioactive signals without incorporating them directly in the system. In that sense, gene therapy and cocultures can be included in this approach. An extensive overview of these approaches is out of the scope of this review. One possible route to overcome this issue is through the use of genetically engineered cell therapy, which has become a cutting-edge approach for tissue regeneration, and it has been under intensive experimental evaluation.^196,197 Cell-based gene delivery approaches to induce bone formation through the transfection of cells with BMP-2, 4, 6, 7, and 9 genes have led to superior bone formation in several animal models.^198–203 Gene therapy approaches for cartilage regeneration have also been explored with transfection of cells with BMP-2, BMP-7, TGF-β1, IGF-1, among others, acting individually or in synergy.^204–208 Blocking angiogenic genes to prevent osteocalcification by using antiangiogenic factors or competitive inhibitors have also shown to promote cartilage regeneration. ²⁰⁹ Menendez et al. ²¹⁰ also assessed osteochondral repair after injection with adenoviral vectors of BMP-2 and BMP-6. Cartilage and subchondral bone regeneration was supported; however, it was insufficient to provide long-term quality osteochondral repair.

Meinel et al. ²¹¹ compared adenovirus gene transfer and protein delivery of BMP-2 on osteogenesis of human MSCs cultured on silk biomaterials. The transfected cells produced BMP-2 within a range of 0–40 ng/mL, and control cultures were supplemented with the same amount of exogenous GF. Results demonstrate that transfection resulted in higher levels of expression of osteogenic marker genes. However, it should be noted that the exogenous BMPs were supplemented in a culture medium, and in our opinion, despite the interesting data obtained with this study, the spatiotemporal control over the release kinetics of the protein should be regarded to obtain a fair comparison between both strategies.

The synergistic effect of combinations of genes encoding for specific GFs and transcription factors on bone and cartilage regeneration has been intensively explored. For bone formation, studies regarding gene combinations of BMP-2 and VEGF, BMP-2, ²¹² VEGF, and Ang-1^213,214; BMP-4 and VEGF^122,215; BMP-7 and PDGF-BB ²¹⁶ ; Runx 2 and Osterix ²¹⁷ ; and FGF-2 and sonic hedgehog ²¹⁸ have been published.

For cartilage regeneration, some examples of combinations that have been used to enhance the formation of new tissue include BMP-2 and IGF-1, ²¹⁹ IGF-1 and IL-1, ²²⁰ FGF-2 and IGF-1,^221,222 Sox5, 6 and 9 genes,^223,224 and TGF-β3 and collagen I silencing short hairpin RNA. ²²⁵ Osteochondral defects have also received particular interest by the combination of BMP-2 and TGF-β1. ²²⁶ Phillips et al. ¹⁸⁵ have also observed graded mineral deposition through the creation of a gradient of Runx2/Cbfa1 oriented along the length of collagen scaffold aiming for orthopedic interfacial TE.

Gene therapy can be applied through two main mechanisms, either the direct delivery of genes into cells and tissues or through transfection of transplanted cells and further seeding onto the construct and/or implantation to the defect site. Both strategies share the vision of cells as factories for bioactive factors. ⁸⁹ The biggest hurdle for the translation of stem cell transplantation into clinical practice has been the in vitro expansion conditions to achieve the required amounts of cells for a successful therapeutic outcome. ²²⁷

The importance of cell-to-cell interactions in the context of TE is also a critical issue. ²²⁸ Cells can stimulate the production of chemoattractant and trophic factors, stimulating neighboring cell populations. During endochondral bone formation, a cascade of signaling occurs between chondrocytes and osteoprogenitor cells that ultimately lead to bone formation via a cartilage intermediate. Chondrocytes can provide morphogenic signals that induce osteogenic differentiation of MSCs. ²²⁹

In development, vascularization precedes osteogenesis, and it is suggested that microvessels accelerate bone formation even before flow has been established. ²³⁰ For example, the coculture of target tissue cells together with endothelial cells is a promising approach to promote vascularization. In the case of bone, considering the intricate connection between angiogenesis and osteogenesis, it is not surprising that the interactions between osteoblasts and endothelial cells happen to be so relevant. Endothelial cells are osteoinductive, as they drive MSCs toward the osteoblastic phenotype. Cocultures of endothelial cells with bone marrow stem cells, osteoblast-like cells, and osteoblast progenitor cells have shown pronounced mineralized matrix production, enhanced microvascular network formation, and increases bone regeneration. ²³⁰

In cartilage, mature chondrocytes are demonstrated to secrete TGF-2, BMP-2, and IGF-1 to direct and enhance MSC chondrogenesis with substantial increase in tissue volume, mass, and ECM production. ²³¹

The special case of PRP

Another promising strategy based on the use of GFs for stimulation of bone, cartilage, and osteochondral healing is through the use of PRP. The implementation of an autologous technology represents a new translational procedure acting as an alternative to the limitations observed with the current TE and regenerative medicine cell-based approaches. ²³² In the past two decades, the increasing knowledge on the physiological roles of platelets in wound healing and tissue injury suggests the potential of using platelets as therapeutic tools. ²³³ Platelets are anucleated cytoplasmatic fragments that form an intracellular storage pool of proteins vital to wound healing. When activated, they release a group of biologically active proteins and other molecules that bind to transmembrane receptors of target cells, leading to the expression of gene sequences that ultimately promote the recruitment, growth, and morphogenesis of cells.^234,235 Most of the GF content is stored in the alpha-granules of the platelets. Platelets activation can be initiated by a number of methods, such as shear forces caused by fluid flow, contact with a variety of materials, including collagen and basement membranes of cells, and thrombin.^236,237 Upon activation of the platelets with thrombin, calcium, or temperature cycles, proteins are released.

PRP can be easily obtained through centrifugation cycles of blood samples, and after activation of the platelet concentrates and consequent liberation of their protein content, the enriched GF cocktail can also be called platelet lysates (PLs). ²³⁴ The standard protocol for the preparation of PRP from autologous blood is based on a two-step centrifugation process to separate blood components into different layers: the separation and concentration steps. For the separation step, blood is centrifuged to separate red blood cells from platelets and plasma. For the concentration step, the supernatant composed of platelets and plasma is collected and centrifuged again to isolate the platelets.^238,239 However, in the literature, several isolation procedures can be found, and they can be divided in three main groups according to the differing parameters in the platelet isolation: (1) final concentration of platelets in PRP; (2) protocol for the activation of PRP; (3) other variations from the standard protocol. ²⁴⁰ There are several classifications to categorize platelet concentrates based on relative concentrations of platelets, leukocytes, and fibrin, but herein the general abbreviation is going to be used.

An attractive approach for the addition of GFs to increase bone, cartilage, and osteochondral regeneration is the addition of PRP to the defect site. ²⁴¹ PRP in the liquid state can be used to disperse and encapsulate cells upon activation of the platelet concentrate, which clots, forms a hydrogel, and creates a 3D environmental support for cells. ²⁴² Moreover, these hydrogels are particularly attractive, because they can be used as minimally invasive injectable systems, with in situ fast gelling abilities. This way, PRP can both act as a GF source and also as a cell carrier for TE applications. Even in strategies where the main role of PRP is the supply of GFs, the presence of the protein concentrate in the structure typically enhances the stability of the scaffold/material and might even work as a glue, as, for example, in constructs build up on the assembly of particles.

The mechanism by which PRP may work has not been fully explored. ²⁴³ However, it is known that the GFs present in the platelet concentrate promote healing by attracting undifferentiated cells into the newly formed matrix and by triggering cell proliferation. ²⁴⁴ Moreover, PRP plays a significant role on the regulation of inflammation, as it may interact with macrophages and limit the degree of inflammation, ²⁴⁵ and on vascularization, as it promotes new capillary growth. ²⁴⁶

There are conflicting reports in the literature with several studies concluding a positive effect of the concentrates in either bone and/or soft tissue healing, while other do not report a significant benefit from the use of the enriched protein suspension for the regeneration of tissues.^247–251 One of the reasons for this disparity is the difference in platelet density used for the different published studies. Platelet count may vary according to the preparation technique, ranging from two- to several fold above the physiological levels.^238,252 To confirm the efficacy of this cocktail of GFs, the isolation procedure should be standardized. ²⁵³ PRP effects change according to its preparation, activation, concentration, protocol for administration, and the material used for the combination. ²⁴⁷ Table 3 summarizes some studies regarding the use of PRP for bone, cartilage, and osteochondral applications. It is clear from the list that there is a pronounced variability in the experimental conditions, specifically on the preparation and isolation of PRP by the suggestion of different centrifugation cycles in time and force.^254,255 Centrifugation force in particular might be a critical step in PRP preparation, as the applied mechanical forces may damage the platelets leading to GF loss. ²⁵⁶

Despite the common extreme variability and donor dependency in the amount of GFs present in the platelet concentrate, typically, the proliferation enhancement is not impaired. ²⁴⁰ However, it has also been reported before that the stimulation effect of the lysates on cell proliferation and differentiation is dose dependent. ²⁵⁷ Previous studies suggest that range of 2-fold to 8-fold increase in platelet concentration above physiological levels of blood samples is required to obtain positive results from PRP. Lower concentrations have suboptimal effects, whereas higher concentrations might have a paradoxically inhibitory effect.^238,258,259 A platelet count of 1 million platelets per μL has also become the benchmark of therapeutic PRP.^260,261

Clearly, PRP also needs to be properly activated to achieve full degranulation of platelets. ²³⁸ The addition of calcium chloride activates PRP, because it replaces the calcium bound by the anticoagulant agent, used to avoid coagulation of the collected whole blood. This calcium allows the conversion of prothrombin to thrombin, thus activating the coagulation cascade. Exogenous thrombin can also be provided, and it has been commonly used as a PRP activator; however, the use of animal-derived thrombin has raised some concerns regarding the potential significant immunogenic and bleeding side effects, including high rates of thrombosis.^243,258 Moreover, some studies reported the loss of osteogenic and chondrogenic potential of platelet concentrate upon activation with thrombin. ²⁷⁹

Studies have also shown that using thrombin as an activator can result in the bolus release of GFs, with nearly 100% of the protein content released within 1 h. Clearly, this method fails to maximize the potential of PRP, as GFs are cleared before they can even promote healing. ²⁷³ Moreover, the need for platelet activation with exogenous thrombin before injection in unclear, since upon PRP injection into connective tissues, it comes in contact with tissue factor, which can activate platelets and initiate the formation of the fibrin 3D matrix.^237,280 To overcome this limitation, other stimulus has been increasingly pursued, especially mechanical destruction through thermic shock.^{240,253,281,282}

Table 3 summarizes some studies promoting the delivery of GFs present in PRP for stimulation of bone, cartilage, and osteochondral regeneration. Kim et al. ²⁸³ evaluated how different activation protocols would affect the concentrations of GFs in the lysates. Curiously, the four activation methods (calcium chloride, a nonionic surfactant, thrombin and calcium chloride, thrombin and calcium chloride with preactivation with shear stress, and collagen) promoted completely different effects on the release of PDGF, TGF-β, FBF-2, and VEGF, and there was no clear pattern on which one of them was more effective. However, the activation method that promoted a higher release of VEGF also stimulated higher bone mineral density and content in critical-size rat craniotomy defects after implantation of β-TCP and the different groups of PRP.

The large list of biological mediators stored in platelets includes the proteins fibrinogen, fibronectin, and vitronectin, which are known to act as cell adhesion molecules. ²⁵⁷ Platelets store essential GF, including PDGF, TGF-β, IGF-1, and EGF.^{238,239,253,284} PRP may act as an exogenous source of TGF-β for bone healing, directing BMSCs to resorption sites. ²⁸⁵ Furthermore, platelets contain different angiogenic factors, such as VEGF, Ang-2, FGF-2, and antiangiogenic proteins, including endostatin and thrombospondin-1, regulating the formation of new blood vessels.^232,239,284

Localized angiogenic factor delivery has proven beneficial for bone regeneration in various animal models by promoting neovascularization, bone turnover, osteoblast migration, and mineralization. Considering that PRP can release factors involved in angiogenesis, platelet concentrates have been used aiming for that goal. Hu et al. ²⁸⁶ reported that PRP possibly starts the angiogenic process by recruiting endothelial cells that line blood vessels and initiates bone regeneration. The angiogenic role of PRP has been reported by Kajikawa et al. ²⁸⁷ and Lyras et al., ²⁸⁸ which, respectively, observed the role of PRP as an activator of circulation-derived cells in the early phase of tendon healing and on early neovascularization enhancement in full-thickness defects of patellar tendon, respectively.

When activated, PLs can even form 3D structures such as hydrogels and scaffolds based on the production of a fibrin network from the fibrinogen released from the platelets and converted to fibrin through the action of thrombin. However, most of the studies report a significant and pronounced volume shrinkage in these structures. ²⁸⁹ It has been stated that the positive role of PRP on bone healing is more related to the fibrin-supporting matrix than for the GFs content. ²⁷⁴ Moreover, the matrix prolongs the exposure of cells to those GFs. On the other hand, it has been stated that PRP enhances osteoprogenitor cell number in the defect area, thus stimulating tissue regeneration. ²⁶¹

Marx et al. ²⁹⁰ proposed the use of PRP to enhance the initial stages of bone wound healing. The GFs and chemokines present in the platelets play critical roles on the chemotaxis, cell proliferation and differentiation, angiogenesis, vascular modeling, and bone formation.^234,252 Bertoldi et al. ²⁴⁷ evaluated the effect of PRP on different stages of bone formation to optimize the administration protocol of the platelet concentrate. They observed that an initial and single dose of PRP was not as effective as the frequent addition of PRP for a long period, which ultimately led to enhanced ALP production and mineralization in osteoblast cultures. These studies support the need for the development of controlled release systems for PRP to enhance tissue regeneration. As an example, Dutra et al. ²⁶⁸ observed superior maturity of bone formation when PRP was associated with bioactive glass foams when compared with nonloaded sponges.

Present research shows that the enrichment with PRP can influence the early stage of bone healing, gradually decreasing the exerted effect with time. ²⁵⁹ It is believed that TGF-β and PDGF promote the healing of soft and bone tissues through stimulation of collagen production to improve wound-healing formation and the initiation of callus formation. ²⁵³ Santo et al.^291,292 also showed that in vitro controlled release of PLs led to a faster osteogenic differentiation of human adipose-derived stem cells, with a stronger contribution during earlier stages in culture.

Several studies reported the effective response of PLs in the repair and regeneration of a variety of tissues other than bone, including cartilage.^293,294 The application of PRP in cartilage repair is relatively new, and there are limited in vivo studies regarding its use for that specific application. However, it has been shown that PRP stimulates chondrocyte and MSC proliferation and cartilage ECM synthesis of proteoglycans and collagen type II. ²⁶⁰ Besides the significant role of TGF-β on bone formation, TGF-β is also one of the most important GFs involved in the process of cartilage regeneration. ²⁹⁵ Wu et al. ²⁶⁶ documented new cartilage tissue in rabbits injected with chondrocytes/PRP mixtures and production of large amounts of proteoglycans and collagen fibrils, whereas Akeda et al. ²⁹⁶ showed enhanced proliferation and proteoglycan and collagen synthesis on porcine chondrocytes. Direct injection of PRP into patient's knee has been increasingly investigated, and preliminary results are promising.^295,297,298

Clinical studies comparing the role of approved BMPs and PRP on bone healing have been done. ²⁹⁹ Calori et al. ²⁹⁹ reported a clinical report regarding the positive role of both BMP-7 and PRP on treatment of long-bone nonunions, with BMP-7 promoting a slightly better healing response in the treated patients. Preclinical evidence has demonstrated that PRP enhances osteoprogenitor cell proliferation, promotes angiogenesis, and enhanced fracture healing and bone regeneration. In a randomized tibial osteotomy trial in 33 patients, ³⁰⁰ the authors reported a superior radiographic and histological healing in defects treated with PRP when compared with the controls, although clinical and functional outcomes did not show significant differences. Considerable effort has also been done to evaluate the role of PRP and PLs as replacement or supplements for in vitro cultures.^{243,295,301–305} The use of animal-derived serum raises concern regarding its immunological response and possible prion and virus transmission and the application of an easily isolated and autologous source of proteins for in vitro cell expansion and differentiation. To translate the culture of MSCs into clinical practice, PLs have also been used to replace fetal bovine serum, thus avoiding the use of animal-derived proteins, showing to accelerate cell expansion, and thus reducing the duration of ex vivo manipulation.^240,282,306

Despite the impressive amount of research with PLs and PRP, a limited number of human clinical trials investigating the use of this concentrate were carried out so far. ²⁵² There is still room for improvement on the use of PRP as a therapeutic agent for skeletal regeneration. The data obtained from studies with humans are mostly obtained from case reports with small sample sizes and in majority from maxillofacial area, with little data regarding the effect of PRP in critical defects in the axial skeleton. ²⁴¹

Moreover, PRP can be used as a useful tool to study the mechanisms underlying the action of several GFs and cytokines. Its widespread availability conjugated with its rich composition provides a highly competent source of signaling molecules. The level of complexity limits the comprehension of the biological phenomenon promoted by PLs; however, there is a potential to maximize this degree of complexity by developing mechanisms of isolating specific GFs from the whole cocktail. Following this approach, PLs could be used not only as a bioactive factor source for tissue regeneration but also as a high-throughput analysis tool for assessing the role of specific molecules released by the platelets.

Combinatory strategies

The latest trends on the application of biomimetic TE approaches require the combination of several biochemical cues, presented through different mechanisms. Therefore, it is expectable that to mimic the complexity of the native ECM, indirect and direct delivery of GFs should be included in the strategy to produce formation of new functional tissue.

The codelivery of PRP with other GFs has also shown high potential. Considering the ability of PRP to stimulate undifferentiated tissue healing mainly through cell proliferation, the combination of this cocktail of GFs with a potent inducing signaling molecule, such as BMP-2 for bone, could lead to a strong enhancement of tissue formation and more importantly to a more directed regeneration pathway toward a specific lineage. The osteogenic potential of BMP-2 and PRP has already been demonstrated.^307,308 It has been demonstrated that PRP reduces the osteoclast-mediated bone collagen degradation, suggesting the inhibition of osteoclast activation. ³⁰⁹ Since BMP-2 stimulates the generation of osteoblasts, but also osteoclasts, the combination of both agents could lead to a more favorable strategy. The combination of gene delivery and direct GF delivery has also been explored. Luo et al. ³¹⁰ evaluated the potential of codelivering VEGF and the gene encoding for BMP-2.

To the best of our knowledge, there are no reports of studies regarding the combination of direct and indirect GF delivery for cartilage and osteochondral interface regeneration.