Advances in Regenerative Orthopaedics

Abstract

Orthopaedic injuries are very common and a source of much misery and economic stress. Several relevant tissues, such as cartilage, meniscus and intra-articular ligaments, do not heal. And even bone, which normally regenerates spontaneously, can fail to mend. The regeneration of orthopaedic tissues requires four key components: cells, morphogenetic signals, scaffolds and an appropriate mechanical environment. Although differentiated cells from the tissue in question can be used, most cellular research focuses on the use mesenchymal stem cells (MSCs). These can be retrieved from many different tissues, and one unresolved question is the degree to which the origin of the cells matters. Embryonic and induced, pluripotential stem cells are also under investigation. Morphogenetic signals are most frequently supplied by individual, recombinant growth factors or native mixtures provided by, for instance, platelet-rich plasma; MSCs are also a rich source of trophic factors. Obstacles to the sustained delivery of individual growth factors can be addressed by gene transfer or smart scaffolds, but we still lack detailed, necessary information on which delivery profiles are needed. Scaffolds may be based upon natural products, synthetic materials, or devitalized extracellular matrix. Strategies to combine these components to regenerate tissue can follow traditional tissue engineering practices, but these are costly, cumbersome and not well suited to treating large numbers of individuals. More expeditious approaches make full use of intrinsic biological processes in vivo to avoid the need for ex vivo expansion of autologous cells and multiple procedures. Clinical translation remains a bottleneck.

“All the successes of orthopaedic surgeons are little more than a reflection of the body’s amazing ability to heal itself”

Henry Mankin, MD¹

Introduction

With its wide range of competencies and frequent injury, the skeletal system provides a rich testing ground for the field of regenerative medicine. The long bones, for example, represent one of the few organs in the human body that normally regenerate spontaneously, without scarring, to restore full function. But large segmental defects in these same bones fail to heal ², and the calvarial bones of the skull lack any inherent reparative ability in individuals older than 2 years of age ³. Likewise, intra-articular ligaments, articular cartilage and menisci have almost no ability to heal after injury. Extra-articular ligaments and tendons, on the other hand, mount a repair response, but produce a regenerate that is inferior to uninjured tissue.

It is not generally appreciated that orthopaedics has long been at the forefront of regenerative medicine. The osteoinductive properties of demineralized bone matrix (DBM), for example, were recognized nearly 50 years ago and it has been available for clinical use for decades ⁴. Two of its active ingredients, bone morphogenetic protein (BMP) -2 and -7, were cloned in the 1980s and have been approved for clinical use since the early 2000s. Cell therapy, in the form of autologous chondrocytes used to heal cartilagenous injuries, has been available since the 1990s ⁵. It should also be pointed out that the concept of a mesenchymal stem cell (MSC) grew out of the orthopaedic research community ⁶, as did the suggestion to use gene therapy for tissue regeneration ⁷. Despite these advantages, progress in developing approved regenerative products for clinical use has been slow.

This review summarizes recent developments and trends in this important area. Space does not permit comprehensive or detailed treatment of the subject matter, but recent reviews are cited for additional information. Emphasis is placed on translational aspects of the field.

Current Concepts

Strategies

Regeneration of the skeletal system, like many other organ systems, requires at a minimum cells, morphogenetic signals and matrices or scaffolds. Because a major function of the skeletal system is to bear load, mechanical stimuli are of additional importance, something that provides an interesting intersection between orthopaedic surgery and rehabilitation.

In the past, major effort was expended in developing traditional tissue engineering technologies for growing new tissue extracorporally prior to implantation (figure 1A). These usually involved harvesting autologous cells from the patient, expanding them in culture, seeding them onto a scaffold and incubating them in a bioreactor. Considerable successes have been logged by this approach ⁸, including the use of patients as their own bioreactors ⁹.

Figure 1. Strategies for regeneration.

Panel A. In traditional tissue engineering approaches, a first procedure removes tissue biopsies. Autologous cells are isolated, seeded on a matrix and incubated in a bioreactor under conditions where new tissue is formed. The graft is implanted into the patient in a second procedure.

Panel B. Expedited approaches require a single percutaneous, minimally invasive, or intra-operative procedure.

Adapted from reference 10

Although technically feasible, the economics and complexities of such approaches mitigate against their widespread application ¹⁰. This point is arguable ¹¹, but there is nevertheless increasing interest in using technologies that do not require ex vivo cultivation of autologous cells for each patient or more than one invasive procedure. This can be achieved with allograft cells, rapid isolation and manipulation techniques that can be used intra-operatively, or by provoking and facilitating endogenous repair processes (Figure 1B). For practical and economic reasons, there is a trend towards minimally invasive approaches.

Cells

Differentiated cells from the tissue to be repaired are already used clinically, as exemplified by autologous chondrocyte implantation (ACI) ⁵. However, attention is increasingly turning to the use of stem or progenitor cells as the basis for skeletal tissue regeneration.

MSCs derived from bone marrow were the first to be investigated, and they remain the cell of choice for many investigators ¹². They can be obtained readily from bone marrow biopsies and easily expanded in culture. Because populations generated in this manner contain cells with the ability to differentiate along various lineages of importance to the skeletal system, they are of wide significance to regenerative orthopaedics. It is commonly believed that MSCs can be successfully allografted, which raises the possibility of facilitating clinical application with universal donor lines. Moreover, MSCs have anti-inflammatory properties of potential benefit when regeneration has to occur in a hostile inflammatory environment.

Initially identified in bone marrow aspirates, MSCs have since been harvested from almost every tissue and organ in the body, including periosteum, long recognized as a rich source of regenerative power in orthopaedics ¹³. Some newer sources, such as sub-dermal fat, are readily accessible and provide far more cells. Fresh lipoaspirates provide a “stromal vascular fraction” comprising, in addition to MSCs, endothelial precursor cells, smooth muscle cells, monocytes, macrophages and lymphocytes, among others. The equivalence between MSCs derived from different sources and their suitability for regenerating different tissues is still under discussion. However, there is an emerging consensus that not all MSC populations are equal, and regeneration is likely to be most successful when MSCs are recovered from the tissue to be regenerated. This is relevant to the suggestion that MSCs from the pulp of extracted human molars, or elsewhere, be banked for future use ¹⁴. Refinements to the use of MSCs include preimplantation sorting to enrich for progenitors of choice, and preconditioning to prime the cells to survive and differentiate, as required, after implantation ¹⁵. A rapidly growing number of companies are developing MSCs for clinical use in regenerative orthopaedics, but at the same time attracting increasing scrutiny by the FDA ¹⁶. Alternative sources of progenitor cells include blood ¹⁷, placenta and umbilical cord ¹⁸. There is increasing interest in the use of endothelial progenitor cells for orthopaedic purposes ¹⁹.

Despite their several advantages, cultures of MSCs display a Hayflick limit and, as they senesce, their differentiation properties alter. Embryonic stem cells (ESCs) and induced pluripotent stem (iPS) cells are of interest because they do not senesce on repeated sub-culture. Moreover, iPS cells obviate the ethical concerns surrounding ESCs. Conditions for their differentiation into cells of the skeletal system are being developed. Safety remains an issue, especially with the potential for teratoma formation.

Morphogenetic Stimuli

Most research has focused on particular growth factors that guide the relevant patterns of differentiation during embryogenesis. Thus BMPs have been widely explored in the context of osteogenesis and transforming growth factor (TGF) -β for chondrogenesis. The use of key transcription factors, such as scleraxis for tenogenesis, is also current. An alternative approach uses small inductive molecules. Patterns of differentiation can also be influenced by the chemical environment, such as hypoxia, which stimulates chondrogenesis.

Delivery of morphogenetic stimuli remains rate limiting for many applications. Although some promising agents, such as teriparatide (parathyroid hormone 1–34; Forteo), can be delivered systemically, for most purposes, it is necessary to deliver the agent locally to the site of regeneration for a period of several days or weeks. This has proved to be extremely difficult to accomplish. Much research thus focuses on the development of smart scaffolds that will provide controlled release of factors at suitable doses for an appropriate period of time. The literature on this topic is voluminous, yet no clinically approved product has appeared. Despite the popularity of this approach, it remains constrained by the reality that we still do not know exactly which factors we need to deliver, at what concentrations, and with what kinetics for individual applications.

Gene delivery offers an alternative strategy for delivering morphogens. In principle, the local delivery of genes (usually cDNAs) encoding the protein in question will result in focal synthesis of a nascent, endogenous product that has undergone authentic post-translational modification. Its synthesis will persist for as long as the cDNA is present and expressed, with the potential for regulated expression. Gene delivery also brings many advantages to the delivery of gene products that work intra-cellularly, such as transcription factors and inhibitory species of RNA; the latter are highly topical, and subject to increasing research. Gene transfer can be accomplished with viral or non-viral vectors, delivered by in vivo or ex vivo protocols; specific examples are given in later sections of this review.

Another option for the delivery of morphogenetic signals follows from the growing realization that MSCs are a rich source of trophic factors. One advantage of this approach is that it is not necessary a priori to know which morphogens are present or required, and the need for their isolation and purification is removed. The same is true of blood derived products such as autologous conditioned serum ²⁰ and the latest orthopaedic panacea, platelet rich plasma (PRP) ²¹.

Scaffolds

Space filling scaffolds provide a substrate, preferably resorbable, upon which new tissue can form. Increasingly, the scaffold is not a passive conductive presence, but an agent that provides cells and signals to induce and inform the development of new tissue. Such signals may be biological, chemical or physical. Agents may be attached to scaffolds covalently or non-covalently by, for example, adsorption, encapsulation, or affinity binding ²². Two general delivery strategies exist. One provides agents at a uniform rate governed, for instance, by the rate of scaffold dissolution. Stimulus-responsive (“smart”) scaffolds deliver in response to changes in the local environment or in response to external stimuli, such as ultrasound or magnetic fields. Scaffolds can also be designed to resorb at rates that coordinate with tissue regeneration. This is achieved by including domains that are susceptible to lytic enzymes produced selectively by tissues during regeneration ²³. Some scaffolds have intrinsic morphogenetic influence, such as those ceramics that, in addition to providing mechanical strength, are also osteoinductive, thus serving a dual mechanical and biological role ²⁴.

Devitalized tissues often provide scaffolds that contain a rich cocktail of endogenous morphogens with inductive properties. A large number are commercially available and approved for clinical use ²⁵; DMB ⁴ is an early example of this. Such materials are versatile, supplied as sheets, powders and gels, and may be shaped to need. Scaffolds may also be associated with gene delivery vectors to form gene-activated matrices (GAMs) that both provide support and deliver cDNA molecules ²⁶.

A wide range of natural and synthetic materials has been evaluated as potential scaffolds (Table 1) and several are in clinical use. The choice of scaffold depends on a variety of factors, including porosity, biocompatibility, resorption rate and mechanical strength. One strategy selects mechanical properties to match those of the tissue being regenerated, so that, as the scaffold resorbs, load is progressively transferred to the new tissue. One disadvantage of this approach is the phenomenon of stress shielding whereby load bearing by the scaffold reduces mechanical stimulation (see next section) of the regenerate.

TABLE 1.

Materials used to form scaffolds for the regeneration of orthopaedic tissues.

Class	Examples
Natural polymers	Collagen Gelatin Silk Fibrin Alginate Chitosan Hyaluronan Coral (hydroxyapatite)
Ceramics	Hydroxyapatite β-Tricalcium phosphate (TCP) Biphasic calcium phosphate (BCP) Calcium sulfate (plaster of Paris) Octacalcium phosphate Bioglass
Synthetic, biodegradable polymers	Polylactic acid Polyglycolic acid (vicryl) Poly (lactic-co-glycolic acid) Polycaprolactone Polyhydroxyalkanoate Polyurethane
Tissue extracellular matrix	Demineralized bone matrix (DBM) Small intestine sub-mucosa (SIS) Skin Dermis Fascia Pericardium
Other	Self-assembling peptides Hybrid scaffolds  Collagen-glycosaminoglycan  Collagen-hydroxyapatite  Gelatin - hyaluronan  Hyaluronan – polycaprolactone  Polycaprolactone - polyurethane

In recent years advanced fabrication technologies have been adapted for scaffold production. Electrospinning ²⁷ allows the formation of nanofibers that can provide complex, anisotropic, biomimetic structures. Advanced imaging and fabrication techniques allow designer scaffolds, anatomically shaped to individual patients ⁸.

Although we usually think of scaffolds as solid implants, there is growing interest in using injectable liquids or pastes that can be applied in a minimally invasive fashion.

Mechanical Stimuli

Mechanical stimulation is essential to the development and function of skeletal tissues. Moreover, it has been appreciated since the 1960s that the mechanical environment can determine stem cell fate ²⁸. Recent research that defines this interaction in greater mechanistic detail ²⁹ promises to provide major advances. The question is how to incorporate this new information into improved regenerative strategies.

Many tissue engineering protocols use bioreactors to provide the appropriate forces ⁸. Control of the mechanical environment is more difficult when attempting in situ regeneration. In the case of bone healing, this has traditionally been addressed with fixtors of different designs. For accessible joints, such as the knee, it is possible to use passive motion machines ³⁰. However, for many applications, control over the mechanical environment is rudimentary. This would seem to be a fruitful area for innovative rehabilitation protocols.

Progress

Bone

For an organ that often heals by itself, bone is proving surprisingly difficult to regenerate. Much present research is focused in four areas: fractures that fail to heal (non-unions), spine fusion, calvarial defects and critical size defects in long bones. The concept of a critical size is based upon the observation that segmental defects beyond a certain size will not heal spontaneously, even in healthy individuals who would otherwise heal a smaller defect ². Much of the research uses young, healthy rodents and rabbits as model systems where osseous defects tend to heal readily. The challenge is to develop technologies robust enough to perform in demanding models using mature large animals, thereby justifying human trials ².

Urist’s observation that DBM has osteoinductive properties ³¹, led to its development as the first commercial product for bone regeneration. This product has stood the test of time and remains in wide clinical use for spine fusion, fracture healing, the repair of cysts, and periodontal reconstruction. ⁴. However, it has several shortcomings including batch-to-batch variability in osteoinductivity. Further research identified a number of morphogens associated with DBM that helped account for its osteoinductive properties. The recombinant forms of two of these, BMP-2 and BMP-7, are approved for specific human applications as the active ingredients of Infuse® (US) (InductOs in Europe) and OP-1® (Ossigraft), respectively. In 2005, the FDA approved the use of platelet-derived growth factor (PDGF) on a tricalcium phosphate (TCP) matrix (Gem21s) for periodontal regeneration. It is presently under review for approval in ankle fusions, although concern about the possible tumorigenicity of PDGF is delaying authorization. Additional factors, such as Nel 1 ³² and wnt ³³ are undergoing pre-clinical development. The latter is highly topical because antibodies against its major physiological antagonist, sclerostin, are coming into the clinic for treating osteoporosis. These antibodies are already in clinical trials for fracture healing (NCT00907296; NCT01081678). Parathyroid hormone 1–34 (teriparatide; Forteo) is clinically approved for treating osteoporosis and has potential use in bone regeneration ³⁴; nine clinical trials are listed in this context on clinicaltrials.gov (visited 3/8/13). Despite current high enthusiasm for biological triggers of regeneration, interest remains in using small molecules ³⁵. A recent novelty in this area involves the use of an organic salt of vanadium as an insulin mimetic to enhance fracture healing ³⁶

Delivery of BMPs is problematic and the commercial products are simply implanted on collagen scaffolds. These provide only burst release, using very large amounts of the product, raising costs and provoking unwanted side-effects ³⁷. Moreover, BMP-2 and BMP-7 have not proved as dramatically effective in human clinical practice as in certain pre-clinical animal models. Considerable research is directed towards the development of scaffolds that deliver growth factors in a controlled fashion ^38,39. However, there is little information on exactly how much growth factor is needed, when it is needed and for how long. Indeed, it is possible to question the degree to which fine control is necessary. Reichert et al., for instance, recently reported very impressive healing of critical size tibial defects in skeletally mature sheep, a very demanding model, using a TCP-polycaprolactone (PCL) scaffold implanted in conjunction with a paste comprising BMP-7 mixed with type I collagen carrier ⁴⁰. Certain ceramic scaffolds are able to induce bone formation without the need for exogenous cells or growth factors ^24,41. Scaffolds for bone regeneration are reviewed in reference ⁴².

Cell-based approaches to regenerating bone also have a substantial history. The clinical use of unfractionated autologous bone marrow as a source of osteoprogenitors, goes back over 20 years ⁴³. Hernigou and coworkers improved the efficiency of the procedure by enriching for MSCs with a cell sorter ⁴⁴. Their data suggest that the injection of a minimum average of approximately 55,000 osteoprogenitors is required to achieve union. Harvest of osteoprogenitors may be aided by the development of improved recovery devices, such as the “reamer-irrigator-aspirator” ⁴⁵. The cells recovered by this device have proved successful clinically as adjuncts to the healing of difficult nonunions and segmental defects, using dexamethasone to promote their osteogenic differentiation ⁴⁶. In a similar fashion, Kim et al ⁴⁷ used dexamethasone to pre-differentiate marrow-derived MSCs into osteoblasts prior to injection into fracture sites. Osteonecrosis has been also treated clinically with marrow ⁴⁸, marrow-derived MSCs ⁴⁹ or the stromal vascular fraction of fat ⁵⁰. Of these publications, only that of Hernigou and Beaujean ⁴⁸ reported a large study with long-term follow-up. They found that when marrow was administered before collapse of the femoral head, only 9 of 145 hips went on to require total joint replacement. In contrast, 25 of 44 hips required replacement when cells were administered after collapse. The MSC content of the marrow aspirates was measured, and better outcomes were associated with the administration of larger numbers of progenitor cells.

Considerable emphasis is being placed on the use of MSCs to repair large segmental defects. Following promising results in rat ⁵¹ and dog ⁵² models, marrow-derived MSCs were used to treat large segmental defects in 3 human subjects ⁵³. All subjects responded well, and were still clinically successful after 7 years, although the slow resorption of the hydroxyapatite scaffold was of concern ⁵⁴. In a related, dental application, autologous MSCs derived from adipose tissue were used successfully to repair calvarial defects in a 7-year old individual ⁵⁵. Perhaps the most striking example of successful bone regeneration concerns the regeneration of the entire distal phalanx of the human thumb ⁵⁶. This was achieved by implanting autologous, expanded osteoprogenitors from the patient’s periosteum loaded onto a coral (porous hydroxyapatite) scaffold (figure 2). No morphogens or other developmental cues were added.

Figure 2. Regeneration of the human phalanx.

Panel A: Thumb at presentation

Panel B: Thumb at one week covered by a pedical of abdominal skin

Panel C: At six weeks, the thumb was well healed

Panel D: Radiograph of thumb prior to implantation of scaffold seeded with autologous periosteal cells.

Panel E: Six weeks after implantation, the thumb was of normal length

Panel F: Radiograph of implant, six weeks post-implantation

Panel G: At 28 months, patient had good pinch strength (2.3 kg).

Panel H: Radiograph at 28 months showing evidence of remodeling.

From reference 56

Based upon the productive interplay between osteoprogenitor cells and a suitable matrix, a system called “Cellect^™” (DePuy AcroMed, Raynham, MA) was developed and made commercially available for the intra-operative harvest and implantation of marrow-derived MSCs. Marrow aspirates were filtered through a suitable scaffold, such as DMB or collagen-TCP, that preferentially retained the adherent MSCs. The entire structure was then implanted into the osseous lesion. Despite encouraging preliminary data for spine fusion ⁵⁷ and osteonecrosis ⁵⁸, this product was discontinued. Other innovative approaches under development include the use of matrix laid down by MSCs as they undergo osteogenesis in vitro. After devitalizing, the matrix is combined with MSCs that have been treated with an inducer of osteogenesis. Implantation of such a construct showed great efficacy in healing a murine calvarial defect ⁵⁹. An additional type of biological, osteogenic membrane was discovered by accident. After surgeons remove large segments of bone, they will sometimes implant an inert spacer until the time of subsequent, reconstructive surgery. Masquelet and colleagues ⁶⁰ noted that the spacers became surrounded by highly osteogenic membranes, which are now used to aid the regeneration of difficult, large osseous defects. Their high osteoinductivity may reflect an unique combination of osteogenic cells and potent osteogenic factors ⁶¹.

Gene transfer offers an elegant way to combine cell therapy with the delivery of growth factors. As pioneered by Lieberman’s group ⁶², marrow MSCs can be genetically modified with recombinant adenovirus vectors to express large amounts of BMP-2, thereby becoming potent osteogenic cells when implanted into a defect. To expedite this strategy, this team now uses lentivirus vectors in combination with buffy coat cells that can be isolated and transduced intra-operatively ⁶³. Other gene therapy approaches, reviewed in references ^64,65, include the direct injection of vectors carrying osteogenic genes ⁶⁶ and the combination of vectors with scaffolds to create GAMs ²⁶. An innovative application of the GAM concept coats allograft bone with recombinant adeno-associated virus vectors to achieve “allograft revitalization” ⁶⁷.

Another innovation takes advantage of the high osteogenic potential of skeletal muscle, as evidenced by the disease fibrodysplasia ossificans progressiva ⁶⁸ and the prevalence of heterotopic ossification in muscle following blast injuries ⁶⁹ and certain types of surgery ⁷⁰. We have reported very efficient healing of femoral, segmental defects in rats using unprocessed muscle biopsies transduced with adenovirus carrying BMP-2 cDNA ⁷¹. A similar, although less reliable, effect was noted using genetically modified fat. This technology also supported healing of calvarial defects, but less dramatically ⁷². Ripamonti and Roden ⁷³ have used morcellized muscle to improve the healing of calvarial defects in the baboon with TGF-β₂. The osteogenic environment of muscle has also been exploited when using patients as their own bioreactors for growing new bone for jaw reconstruction ^74,75.

Gene transfer technologies can also deliver inhibitory RNA molecules; knockdown of chordin ⁷⁶ and noggin ⁷⁷, two inhibitors of BMPs, enhances the osteogenic differentiation of MSCs.

Bone forms physiologically by two routes. Intramembranous, or direct, osteogenesis occurs when osteoprogenitors differentiate directly into osteoblasts. Endochondral ossification occurs via a cartilagenous intermediate that is replaced by bone. Strategies for the regeneration of bone have tended to focus on the former which has to confront the early need for a blood supply; this explains why vascular endothelial growth factor (VEGF) is a favored growth factor in many studies. However, it has proved very difficult to engineer an effective vascular supply for regenerating bone. Under these conditions, the endochondral route is attractive because it relieves the physician of the need to provide a blood supply ^78,79. Chondrogenesis does not require angiogenesis and the biology of endochondral ossification supplies angiogenic signals spontaneously as the process evolves. Much recent attention is thus focused on the endochondral route to bone regeneration, especially as chondrogenesis is favored by the hypoxic, acidotic conditions of major lesions. It may be particularly useful in regenerating bone at sites, such as the diaphysis of long bones, where the endochondral process is the normal route of healing.

As noted, mechanical signals are important for osteogenesis and clinical research has confirmed that a certain level of micromotion promotes fracture healing. The concept of dynamization has also proved attractive, whereby bones are fixed rigidly to initiate healing and then allowed axial motion to promote maturation and remodeling. We have recently studied this experimentally in a rat, femoral, segmental defect model and found, unexpectedly, that healing is accelerated by first fixing loosely and then imposing high stiffness as bone begins to be deposited. We call this “reverse dynamization” ⁸⁰. Sophisticated manipulation of the mechanical environment provides additional avenues for bone regeneration.

Bone regeneration has been reviewed in references ^81,82.

Cartilage

Cartilage is frequently damaged as a result of sporting injuries or other trauma, and is eroded in joints with arthritis. Damaged cartilage often leads to joint pain and can predispose to osteoarthritis (OA). Cartilage repair is indicated for symptomatic, but otherwise healthy, joints using one of a variety of approaches indicated in this section. Repair is not normally attempted in arthritic joints, because of the size and nature of the lesions, as well as recognition that the concomitant disease process could impair restoration of cartilage. Strategies for articular cartilage regeneration may differ, depending upon whether the lesion is restricted to the cartilage itself, or penetrates the underlying bone to form an osteochondral lesion.

Unlike bone, articular cartilage has almost no intrinsic ability to regenerate. The lack of a repair process is usually ascribed to the low cellularity of cartilage and the absence of blood, lymph or innervation. One way to obviate this is to allow communication between the cartilagenous defect and the underlying marrow by piercing the sub-chondral bone. A clot forms where MSCs from the marrow differentiate and synthesize cartilagenous repair tissue. A simple, arthroscopic procedure, such as microfracture, is often clinically effective, especially for lesions <2cm², even though the repair tissue is a fibrocartilagenous scar rather than true hyaline cartilage. The major reasons for failure are the inferior mechanical properties of the repair tissue and bone invasion.

ACI is an alternative clinical procedure in which healthy cartilage is harvested from a non-weight bearing part of the joint as a source of autologous chondrocytes. These are expanded in monolayer culture and, in its original configuration, reimplanted as a suspension under a flap of periosteum. Despite the phenotypic modulation of the chondrocytes during serial, monolayer passage and the absence of a scaffold, this procedure is surprisingly effective, and produces a clinical response arguably equivalent to microfracture ⁸³, but suitable for larger defects. Improvements to the technology include using a collagen flap to replace the need for periosteum, and seeding the chondrocytes onto a collagen scaffold prior to implantation in a procedure known as MACI (matrix associated chondrocyte implantation). Selection of the most chondrogenic cells prior to implantation may also improve outcomes ⁸⁴.

The major disadvantages of ACI are the need for two invasive procedures and the extensive expansion of cells for each patient. Moreover, there is evidence that harvest of donor cartilage initiates degenerative changes elsewhere in the joint ⁸⁵. Because cartilage is thought to be an immunologically privileged site, there is interest in using allografted chondrocytes, especially from juvenile donors that may be less antigenic ⁸⁶. One such product, DeNovo® NT (Zimmer, Warsaw, IN), is already on the market.

MSCs provide an attractive alternative. Marrow-derived MSCs undergo chondrogenesis in vitro in response to TGF-β, but there is a concern that they will undergo further differentiation to osteoblasts, leading to the formation of bone where there should be cartilage. This property may account for the observation of bone invasion after microfracture. Nevertheless, marrow-derived MSCs have been used successfully in the clinic to repair cartilage lesions ^87,88. O’Driscoll has pioneered the use of periosteum to regenerate cartilage ⁸⁹ and has used this tissue to treat lesions clinically ⁹⁰. The stromal vascular fraction of adipose tissue has also been evaluated in two patients ⁵⁰. A protocol using progenitors from umbilical cord is in clinical trials (NCT01041001).

Archer’s group have identified a population of chondroprogenitor cells that exists in the superficial zone of articular cartilage ⁹¹. These cells undergo chondrogenesis without forming bone and are undergoing pre-clinical evaluation as allografts in large animal models of cartilage repair. Protocols for the chondrogenic differentiation of ESCs ⁹² and iPS cells ⁹³ are being developed.

Because cartilage is a highly hydrated tissue, many of the scaffolds being developed for cartilage repair are hydrogels. Materials include fibrin, hyaluronic acid, collagen, chitosan, silk, alginate and synthetic polymers such as poly lactic acid (PLA) and poly glycolic acid (PGA). Newer hydrogels are based on poly ethylene glycol (PEG), self-assembling peptides and electrospun nanofibers. Devitalized cartilage provides another scaffold of interest ⁹⁴, whose investigation is encouraged by the successful regeneration of a trachea, an organ which contains articular cartilage, from devitalized donor tissue ⁹⁵. Another approach dispenses with matrices altogether and instead allows chondrocytes in culture to develop their own extracellular matrices, creating implantable grafts ⁹⁶ of the type exemplified by the DeNovo® NT product mentioned above.

Biological and mechanical integration of engineered cartilage grafts in articular cartilage is problematic ⁹⁷, exacerbated by the fact that, unlike bone, articular cartilage does not remodel rapidly or extensively. One strategy for improving biological integration involves local digestion of the surrounding cartilage using enzymes or catabolic cytokines ⁹⁸. Mechanical integration may be helped by recent research suggesting that exposure of neo-cartilage to a combination of fibroblast growth factor-2 and TGF-β promotes rapid maturation with enhanced mechanical properties ⁹⁹. Another approach subjects chondrocytes to hydrostatic pressure in a bioreactor to accelerate maturation prior to implantation ¹⁰⁰. This product is being evaluated in a Phase III human protocol (NCT01066702).

Rather than implant rigid, pre-formed tissue grafts, there is interest in applying soluble materials that can take on the shape of the defect and then solidify in situ. This leads to greater filling of the defect and enhanced integration with surrounding cartilage. One approach uses a hyaluronan-based polymer that is liquid at room temperature and solidifies at body temperature ¹⁰¹. A recent publication describes a polymer based on PEG that solidifies in response to light. It has been used successfully in a pilot clinical study (figure 3) to improve the outcome of microfracture by providing a reliable matrix for colonization and chondrogenesis by MSCs entering from the marrow ¹⁰². In rabbits, entry of MSCs into defects can be increased by incorporation of the chemotactic protein, stromal cell-derived factor 1 ¹⁰³.

Figure 3. Clinical procedure for adhesive implantation into a cartilage defect.

Panel A: A mini-incision exposed the cartilage defect.

Panel B: The adhesive was applied to the base and walls of the defect followed by surgical microfracture.

Panel C: The hydrogel solution was injected into the defect and photopolymerized in situ with light.

Panel D: Bleeding from the microfracture holes was trapped in and around the hydrogel.

From reference 102

Our group has combined this type of approach with gene transfer. Marrow was harvested from rabbits, mixed with recombinant adenovirus vectors, and allowed to clot thus forming a “gene plug”. The soft clot could be press-fit into full thickness defects, with the adenovirus vector transferring genes to MSCs as they entered from the marrow ¹⁰⁴. The high efficiency of gene transfer using adenovirus vectors associated with hydrogels, including fibrin, helps explain the success of this technology ¹⁰⁵. Promising data were subsequently obtained using this method to deliver TGF-β in a sheep, chondral defect model ¹⁰⁶.

The ex vivo transfer of genetically modified chondrocytes ^107,108 could be piggy-backed on standard ACI procedures. Indeed, a clinical trial is about to start in Korea in which retrovirally transduced chondrocytes that over-express TGF-β will be implanted as allografts into lesions in cartilage (Bumsup Lee, personal communication).

Gene therapy could also be used in conjunction with MSCs from various sources ^109,110. Preliminary data suggest that genetically modified fat and muscle grafts might also be effective ⁷¹. Madry’s group has pioneered the direct application of recombinant adeno-associated virus to sites of articular cartilage defect, with encouraging results in animal models ¹¹¹.

Gene therapy approaches are reviewed by Cucchiarini and Madry ¹¹².

The discussion so far refers to the healing of small to medium sized lesions in cartilage. As the field of cartilage regeneration develops, it confronts the need to restore very large lesions or even resurface entire joints. Mao’s group succeeded in resurfacing the rabbit proximal humerus using an anatomically correct, implantable scaffold of PCL and hydroxyapatite in conjunction with a collagen hydrogel containing TGF-β₃ ¹¹³. The data suggest that TGF-β₃ both recruited chondroprogenitor cells, possibly from the synovium, and promoted their differentiation into chondrocytes.

Repair of cartilage defects in OA is more challenging because there is a concomitant disease process creating an unfavorable repair environment. For example, our group has shown that inflammatory cytokines strongly inhibit the chondrogenic differentiation of MSCs ¹¹⁴. Moreover, there are data to suggest that marrow-derived MSCs from individuals with OA are intrinsically less chondrogenic ¹¹⁵. One approach to regenerating cartilage in joints with OA, already in clinical trials (NCT01671072), involves the intraarticular injection of genetically modified, allogeneic chondrocytes expressing elevated amounts of TGF-β₁ ¹¹⁶.

The influences of mechanical forces on chondrogenesis are widely appreciated, but incompletely understood ¹¹⁷. Recent data suggest that shear forces are important in this regard ¹¹⁸, promoting the endogenous production of TGF-β These are the sorts of forces that could be produced by a modified continual passive motion machine of the type already used in rehabilitation after knee surgery ³⁰.

Cartilage regeneration has been review by ^119,120.

Other organs

Intervertebral disc

The IVD is a major load bearing structure of the spine, and its degeneration is associated with back pain. This is currently treated surgically by removing the degenerate disc and fusing the adjacent vertebrae, often using BMP-2 to enhance the process. Prosthetic discs have been developed, but are not a clinical success. Attempts to regenerate the disc have to accomodate the anatomy of the disc, where a highly hydrated, gelatinous nucleus pulposus exists within a fibrous, collagenous annulus fibrosis. These two structures differ considerably in biology and mechanical properties ¹²¹.

Attempts to regenerate the nucleus pulposus by the intra-discal injection of growth factors has met with some success in animal models, and the FDA has given permission for clinical trials using BMP-7 and growth and differentiation factor -5 (GDF-5; also BMP-14) in this fashion ¹²². Recognizing the probably transient effect of injected growth factors, there is interest in using gene transfer to provide sustained delivery ¹²³. As the disc is so physiologically isolated and cell turnover is low, it is possible to obtain remarkably long periods of transgene expression, even using highly antigenic vectors such as adenovirus.

Cell therapy is attractive because IVD degeneration is associated with cell death; both disc cells and MSCs have been evaluated in this regard. Because the nucleus pulposus of the disc is highly acidotic and hypoxic there are concerns about the survival of transplanted cells. Preconditioning has been suggested to prepare cells for this environment and help engraftment. In a small clinical trial, suspensions of autologous, expanded, nucleus pulposus cells were injected into discs following surgery for disc prolapse. Preliminary data are encouraging ¹²⁴. Yoshokawa et al. reported the intra-discal injection of autologous MSCs in two patients ¹²⁵. The discovery of progenitor cells within the disc ^126,127 adds new possibilities for this type of therapy. One impediment to research into cellular therapies for IVD regeneration is the lack of good markers for the relevant disc cells.

Survival and function of cells could be aided by a suitable scaffold, and a variety of hydrogels based upon chitosan, hyaluronan, alginate, cellulose, and composites of collagen/hyaluronan, and chondroitin sulfate have been investigated. The annulus fibrous has much greater tensile strength and materials explored in annulus regeneration include PLA, poly octanedial malate, gelatin, silk, and poly caprolactone triol malate. Electrospun PLA has been examined as a way of forming the alternating, lamellar structure of the annulus.

As noted earlier in the context of OA, degeneration of the IVD is often part of a disease process that needs to be controlled to allow successful regeneration. The research of LeMaitre and colleagues ¹²⁸ makes a strong case for the involvement of interleukin-1 in IVD degeneration. Moreover, IVD degeneration is associated with calcification of end plates, which limits diffusion. Another limitation to progress in this area is the lack of good animal models ¹²⁹.

Disc regeneration has been reviewed in references ^121,130.

Meniscus

The menisci are responsible for load transmission through the knee and are frequently injured. At one time it was common to remove the offending menisci; realization that this pre-disposed to OA spurred efforts to repair or regenerate menisci. The outer third of the meniscus (the red zone) is attached to the synovium and has a blood supply, providing it with some potential for repair. The inner two thirds (the white zone) lack a blood supply and cannot regenerate.

The implantation of a cell-free, collagen scaffold is already used clinically with highly promising results ¹³¹. Scaffolds based upon polyurethane are in clinical trials ¹³² and a product based on silk is expected to launch this year. Pre-clinical research is evaluating alternative scaffolds, such as poly (lactic-co-glycolic acid) (PLGA), hyaluronan-PCL, and devitalized meniscus, augmented with VEGF or PRP or cells. Types of cells include MSCs, and autologous or allogeneic meniscus cells. A provocative paper by Murphy et al. ¹³³ reporting the regeneration of the meniscus in goats following the injection of marrow-derived, autologous MSCs suspended in hyaluronan, has prompted two clinical trials (NCT00225095; NCT00702741).

A small number of studies (e.g. ^134–136) have explored the use of gene therapy

Approaches to meniscus regeneration have been catalogued by Pereira et al. ¹³⁷.

Ligament and Tendon

Although ligaments and tendons are different tissues, it is convenient to combine their discussion in a short review. Most pre-clinical research in this area uses injuries to the anterior cruciate ligament, Achilles tendon or rotator cuff as experimental models.

Two types of regenerative requirements exist, depending upon whether the need is to regenerate tissue in the center of the structure, or the insertion site to the bone or tendon-muscle junction. The former is simplified by the relative homogeneity of the surrounding tissue, whereas the insertion and junction sites are complex, multi-tissue, graded entities ¹³⁸.

Because ligaments and tendons are collagenous, there is much interest in using growth factors that promote collagen synthesis, including TGF-β, PDGF, and insulin-like growth factors, as well as autologous conditioned serum ²⁰ and the ubiquitous PRP. BMPs -12, -13, and -14 (GDFs -7, -6, -5) ¹³⁹, and the transcription factor scleraxis ¹⁴⁰, are of particular interest because they promote the differentiation of progenitor cells into tenocytes and ligament cells. Because scleraxis is an intracellular protein, gene transfer is a useful modality for its delivery that has been used successfully to promote tenogenesis in vitro ^141,142. Other gene therapy strategies include various viral and non-viral vectors used in vivo and ex vivo (summarized within reference ¹⁴³).

Most cell-based approaches to regeneration have adopted ex vivo strategies using tenocytes, ligament cells and skin fibroblasts ¹⁴⁴. However, both tendons and ligaments have been found to possess local populations of stem cells ^145,146, raising the possibility of stimulating endogenous repair processes. Collagen, synthetic polymers such as PLGA, and devitalized extracellular matrix preparations are frequently the scaffold of choice (reviewed in reference ¹⁴⁷).

Mechanical forces are important in the formation and maturation of ligament and tendon; tensile stress, for instance, upregulates the expression of scleraxis by MSCs ¹⁴⁸. There is interest in harnessing mechanobiological processes to regenerate the region where tendon or ligament inserts into bone. This is challenging, because this includes a fibrocartilagenous intermediary zone. An alternative approach uses multiphasic scaffolds containing multiple cells types. In one example, scaffold was seeded with ligament cells at one end, osteoblasts at the other, and chondrocytes, in between ¹⁴⁹. A more practical approach may be to use a single progenitor cell type and induce zone-specific differentiation by spatial differences in the matrix ¹⁵⁰.

Ma et al ¹⁵¹ recently engineered a scaffold-free, bone-ligament-bone anterior cruciate ligament graft in vitro using marrow-derived MSCs from sheep. Six months after implantation into ovine knee joints, the engineered structure had attained much of the biology and mechanical properties of the native, sheep ACL.

Unlike the case with the other tissues described in this review, there have been no major clinical trials adopting the strategies of regenerative orthopaedics for healing ligaments and tendons. However, a variety of scaffolds derived from devitalized extracellular matrix are used clinically to augment repair.¹⁵²

Perspectives

As reflected in the opening quotation, biology provides many of the tools needed to restore orthopaedic tissues, and our research can be viewed as attempts to harness these tools to achieve regeneration in clinical settings. This review indicates strategies for achieving this goal. Despite many pre-clinical successes, there are few clinical trials and even fewer new products for the regeneration of orthopaedic tissues are entering clinical use. The reasons for this are a matter of discussion, ranging from regulatory burdens, funding restrictions, gaps in scientific knowledge, inadequate animal models and psychosocial issues ¹⁵³. Health care economics play an increasingly important role, which is one reason, among others, that this author favors approaches that facilitate endogenous repair ¹⁰. The key scientific components of success appear to be in place and it would seem only a matter of time before a variety of new products are available to address the growing need for effective and affordable regeneration of skeletal structures.

Abbreviations

ACI

Autolgous chondrocyte implantation

BMP

Bone morphogenetic protein

DBM

Demineralized bone matrix

ESC

Embryonic stem cell

GAM

Gene activated matrix

GDF

Growth and differentiation factor

iPS

Induced, pleuripotent stem

IVD

Intervertebral disc

MSC

Mesenchymal stem cell

OA

Osteoarthritis

PCL

Poly caprolactone

PDGF

Platelet-derived growth factor

PEG

Poly (ethylene glycol)

PGA

Poly glycolic acid

PLA

Poly lactic acid

PLGA

Poly (lactate-co-glycolate)

PRP

Platelet rich plasma

TCP

Tricalcium phosphate

TGF

Transforming growth factor

VEGF

Vascular endothelial growth factor