The knee is a multi-component organ system comprised of several tissues which function coordinately to provide mobility. Injury to any one component compromises the integrity of the system and leads to adaptation of the other components. Over time, such events often lead to dysfunction and degeneration of the knee. Therefore, there has been considerable research emphasis to repair injured components in the knee including cartilage, menisci, and ligaments. Approaches to improving healing and repair/regeneration of knee tissues have included surgery, anti-sense gene therapy, injection of growth factors and inflammatory cytokine antagonists, transplantation of in vitro expanded chondrocytes, enhancement of endogenous cells via microfracture, injection of mesenchymal stem cells, and implantation of in vitro tissue engineered constructs. Some of these approaches have lead to temporary improvement in knee functioning, while others offer the potential to restore function and tissue integrity for longer periods of time. This article will review the status of many of these approaches, and provide a perspective on their limitations and potential to contribute to restoration of knee function across the lifespan.
The knee is a complex “organ system” which is comprised of several tissues (e.g. intraand extra-articular ligaments, tendons, muscle, cartilage, menisci, capsule, and synovium) which arises during development to function as an integrated unit during mobility [1-3]. All of the tissues of the knee are vascularized and innervated to variable extents except the articular cartilage which is devoid of vascularity and innervation, likely in part due to its requirement to function in a very intensive biomechanical loading environment.
Following development during fetal life, the tissues of the knee are required to proceed through post-natal life with maturation, puberty and finally skeletal maturity steps, while maintaining their coordinate functioning [
The major responsibility to adequately maintain the integrity of the extracellular matrix matrix (ECM) in the various tissues of the knee, which is critical for the functioning of the knee falls to a variety of cells that populate the tissues. Each tissue is populated by a variety of cell types uniquely differentiated to both survive in specific biomechanical environments, and to provide contributions to load-dependent tissue homeostasis [
Not only do cells in the tissues respond in a positive manner to biomechanical stimulation by increased expression of anabolic genes (e.g. growth factors, ECM components), but also they respond to a lack of mechanical stimulation/immobilization [discussed in 1,4,11]. That is, tissues of the knee subscribe to the “Use it or lose it” paradigm. If the tissues experience a decline or loss of biomechanical stimulation, they induce a set of catabolic genes that contribute to atrophy of the tissue down to a new homeostatic “set point” (e.g. the point at which a balance between anabolism and catabolism is reestablished for the new mechanical demands) where this catabolic subset of genes is again repressed [
Such findings have implications for both approaches to enhancing repair/regeneration using enhancement of endogenous cell function and associated mechanisms, as well as exogenous approaches to tissueing engineering of tissue replacements. An example of the latter is the work of Goulet and co-workers who have shown that biomechanical loading of a tissue engineered ACL replacement leads to a stronger more functional engineered tissue when transplanted in vivo into goats or dogs [reviewed in 12].
In most tissues of the knee (the aneural, avascular articular cartilage is a notable exception), the cells of either the injured tissue, or puripotent/multipotent (e.g. mesenchymal stem cells [MSC] or mesenchymal progenitor cells [MPC]) cells associated with the synovial fluid or synovial membrane of the intra-articular environment are available for assisting in the healing process [discussed in 13-15].
Following injury to one or more tissues of the knee, or development of a degenerative condition (e.g. idiopathic with no known overt injury) such as osteoarthritis, leads to a loss of joint integrity. Injury to some tissues, such as the extra-articular MCL, leads to the functional healingbut slow remodeling of the tissue in preclinical models [reviewed in 16]. For the MCL, a complete transection or rupture of the tissue in most instances does not require surgical intervention and the healing proceeds normally. The initial phases of wound healing (inflammation, matrix deposition, and slow remodeling) take place but the matrix deposited is not normal and is a scar-like tissue. While the cellularity of the healing MCL gradually returns to normal levels, the biomechanical integrity of the healed MCL is somewhat compromised up to 2 years later [
Use of in vitro expanded autologous and allogenic chondrocytes and meniscal cells, and subsequent implantation in vivo have been extensively studied over the past decade or more [reviewed in 20-25]. While somewhat expensive, this approach of expanding endogenous cells has proven to be successful for a number of patients, particularly younger patients with limited defects in their articular cartilage of the knee. However, removal of autologous cartilage from non-weight bearing areas of the knee cartilage to obtain chondroctyes does induce some damage to the remaining cartilage in the knee, and there are some concerns regarding dedifferentiation of chondrocytes in vitro during expansion. In spite of limitations, this approach appears to have fairly long term clinical efficacy, particularly for a subset of patients.
Attempts to enhance endogenous repair (which is not regeneration) via approaches such as treatment of early scar tissue with anti-sense reagents to specific molecular targets were only partially successful, but certainly proved the concept was potentially viable [26,27]. In a rabbit model, the injured MCL was injected with either an antisense reagent directed towards the small leucine-rich proteoglycan (SLRP) decorin, a sense reagent, or vehicle early after injury. Subsequent sacrifice and analysis revealed that the anti-sense treated MCL was biomechanically stronger than the others, and that significantly larger numbers of collagen fibrils were detected in the antisense treated animals [
A second approach to enhance the activity of endogenous cells is via the use of reagents such as Platelet-Rich Plasma (PRP) [reviewed in 30-34]. As normal platelets contain a number of mediators, including growth factors and other biologically active molecules, injection of such preparations could result in the release of several “repair-enhancing” molecules or molecules which may negate the impact of endogenous mediators which are exerting a negative or counter-productive influence on repair. Since such preparations can be used in an autologous manner, this is a significant advantage. However, being autologous can also be a disadvantage if an individual has compromised platelets or plasma.
While the rationale for using such platelet preparations is attractive, the literature would indicate that responses to injection of such preparations is quite variable (see reviews indicated above), and there are not a large number of appropriately controlled clinical trials. Additionally, the outcome measures used by different investigators in various studies can differ, and range from those focused on actual repair to symptomatic relief (most notably pain). As platelets, and their contents, can be influenced by a number of factors, it is not surprising that outcomes may also be variable. Furthermore, the disease setting (early, late; concurrent use of medications, etc.) also can impact effectiveness. Thus, for PRP, the “jury” is still out regarding effectiveness and the mechanisms by which it is effective in some individuals, and whether it is more effective in acute (e.g. torn menisci, tendons, ligaments) versus chronic conditions (e.g. tendinosis, osteoarthritis with cartilage degeneration). Likewise, when PRP is “successful” it is not known whether it is influencing endogenous fibroblasts, chondrocytes, fibrochondrocytes, or is actually working at the level of mesenchymal stem/ progenitor cells in the damaged tissues, or derived from a nearby source such as the synovial fluid or synovial membrane.
It is clear that preparations such as PRP contain multiple growth factors and other mediators, and the effectiveness is variable. At the opposite end of that scale, a number of studies have attempted to enhance repair of specific joint tissues, or restore joint function via use of individual growth factors, or reagents known to influence the activity of molecules within the abnormal joint. For the most part, such studies with growth factors known to be involved in repair have not exerted a highly significant impact on acute repair in normal healthy individuals/preclinical models [e.g. 35]. However, a study from Salo’s group [
Other researchers have taken a somewhat different tack to interfere with abnormal processes either within an injured joint or fibrotic processes in general. An example is studies which showed that treatment of mice knees with anti-TGF-beta antibodies inhibited osteophyte formation in mouse knees with osteoarthritis [
An additional approach to enhance restoration of function by endogenous cells has been developed by orthopedic surgeons in an attempt to restore cartilage loss. This procedure involves micro fracturing of the bone that was exposed by articular cartilage loss to induce blood, bone marrow elements, and likely stem cells associated with the subchondral bone to localize and form a fibrocartilage-like surface in the compromised area [reviewed in 41-43]. As the fibrocartilage formed by such a procedure is not as resilient as the hyaline cartilage that was there initially, this procedure is likely temporary although some patients experience prolonged benefit [
In summary, attempts to enhance functional and structural repair by stimulating endogenous cells in the damaged tissues or the processes known to stimulate wound healing have met with mixed results. The findings have been very dependent on variability in individual responses, or effects have been dependent on local environments. This latter point is likely very relevant to some of the discussion which follows related to applications of stem/ progenitor populations for cellular repair, as well as tissue engineered tissue replacements.
It is clear from the literature that many tissues contain cells which have some properties consistent with them being either mesenchymal stem or progenitor cells. In addition, such cells can be readily found/detected in bone marrow (BM), adipose tissue (peripheral fat obtained by liposuction or fat pads within the knee) [reviewed in 44], skin, synovial membranes (SM), as well as synovial fluid (SF) [45,46; and many others]. While such cells appear to be able to differentiate into different lineages (e.g. bone, cartilage, adipocytes, fibroblasts), cells obtained from BM appear to preferentially become bone-related, while those of SM and SF appear to be oriented towards chondrogenesis [47,48]. Such variation, plus ease of obtaining the cells, can influence the approaches taken to address either approach: in vitro tissue engineering versus injection of cells into the knee and hope they home to the damaged area and initiate repair. These two approaches are intrinsically different, but have to meet some common challenges when one considers joint restoration from a structural and biomechanical perspective versus injection of cells which need to home to the damaged tissue and initiate repair, often times in the absence of a suitable scaffold such as in advanced osteoarthritis with cartilage loss and/or tears and degeneration of the menisci.
As discussed above, mesenchymal stem/progenitor cells (MSC/P cells) have been isolated repeatedly from a number of tissues/sources that would permit autologous use. In addition, cells with such characteristics have also been isolated from placenta and related tissues [49,50] which could possibly be used in an allogeneic setting. While many of the studies employing isolated cells have used preclinical models to characterize effectiveness, others have used such cells in patients with osteoarthritis and other knee conditions. Unfortunately, in some of the latter situations, their use has been empirical and not supported by adequately controlled clinical trials [discussed in 22,51; and many others]. However, this has been changing and for just adipose-derived cells, over 40 clinical trials have been registered with the NIH (USA) by 2012 [discussed in 44]. Furthermore, in many cases, the outcomes are related to symptom (e.g. pain) management rather than tissue repair and structural restoration, but likely this will also change going forward as more clinical trials come to fruition. Therefore, in the coming few years, evidence for the efficacy of MSC from various sources will be available and informed decisions on their use will be possible.
One of the key findings to date is that only a small percentage of the MSC populations injected into a joint actually remain in the joint or at the site of the injury. For instance, it was shown that very few labeled injected cells associated with induced defects in knee menisci [52; and many others]. This is a common finding, that following injection of literally millions of cells, only a small percentage of cells remain at the site of injury over time. There are likely several explanations for this finding. Firstly, populations of cells with the phenotype of MSC are quite heterogeneous, even from a location such as synovial fluid, with respect to proliferation and differentiation potential [48; Kutcher et al., in preparation; J. Kutcher, MSc Thesis, University of Calgary, 2012]. Isolating cloned cells by limiting dilution analysis of synovial fluid MSC has revealed significant heterogeneity in the characteristics of such cells. Thus, perhaps only a small percentage of what are called MSC from a particular source, are likely actually able to home to a site of injury and contribute to repair of the injury either directly, or indirectly via secretion of essential growth mediators or molecules facilitating repair (e.g. “nurse cells”). The observations after injection of MSC into knees are not unique to this environment, and this outcome has also been observed in models of heart damage and other tissue injuries.
A second potential complicating factor associated with injecting “normal” MSC into injured or damaged knees is that the inflammatory environment resulting from the injury is inhospitable for the proper functioning of the injected MSC. MSC isolated from the synovial fluid or synovial membranes of sheep knees that had incurred an injury led to lower ability of the cell population to differentiate towards the chondrocyte phenotype [
Analogous studies with human knee MSC derived from SF of normal knees, and knees from individuals with osteoarthritis (OA) have provided further support for the concept that inflammation in a knee arising from either an injury or a disease such as OA can lead to modifications in the effectiveness of the cells. Firstly, MSC from OA patients apparently have lost their ability to selfaggregate and differentiate towards a chondrogenic phenotype [
Based on the above discussion, it may be critical to minimize the inflammation in the joint prior to injecting MSC to facilitate repair. This is likely true for both treatment of early OA, or alternatively, repair of knees subjected to inflammatory arthritis such as rheumatoid arthritis [
Finally, injecting MSC from a variety of tissue sources has also been attempted where the outcomes are not related primarily to structural repair of the tissues (although some structural improvement has been noted in some patients), but instead, the alleviation of symptoms (e.g. primarily pain such as in OA). This is not unreasonable as stem cells from some sources have immunomodulatory properties [56,57; and many others]. Thus, injecting MSC into knees with advanced OA does not offer much hope to completely repair the extensive damage to the articular cartilage, menisci, and other involved tissues, but if the MSC were indeed a supply of anti-inflammatory molecules, or molecules that may influence neural activities contributing to the pain of OA, then they may have value alone or in conjunction with pharmaceuticals. As an example, injection of MSC (such as obtained by liposuction or bone marrow aspiration) has been reported to alleviate pain in some patients [58-60; and many others]. They may not be effective in this regard in all patients, but certainly many patients report receiving benefit. Whether some of such benefit is related to the effectiveness of the cells, or is due to a “placebo” effect is not clear, and one has to consider the placebo effect as a factor with respect to symptoms such as pain which can be subjective. This point stresses the need for well controlled clinical trials in this area to tease out where the benefits are and to perhaps identify patient subsets that would benefit more than others from injections of such cells (e.g. precision or personalized medicine).
From the above discussion regarding cells, it is very challenging to repair an extensively injured or damaged knee with multiple tissues involved and with respect to articular cartilage, complete loss of tissue resulting in bone-on-bone and very little residual scaffold to use as a template. Therefore, likely the best application for in vitro tissue engineering of tissue replacements will be for limited size defects in the articular cartilage or menisci, or for a tissue of defined specifications [e.g. an ACL replacement; 61].
For the latter, many studies have used fibroblasts rather than MSC [discussed in 62; and reviewed in 12, 63]. However, such in vitro engineered constructs (boneligament-bone configurations) have been implanted into goats and they have survived and functioned. Whether these avascular, aneural constructs gradually become more “scar-like” over time as host cells infiltrate the implanted engineered construct remains to be confirmed. A key aspect of generating such engineered constructs of fibroblasts + a scaffold is the application of mechanical loading during the generation of the construct. Such loading appears to lead to a much stronger and organized construct which then has a much better chance for survival once implanted as a “reconstructed” ACL even if it is a single band of ligament-like collagen-based material. In the future, it is likely that the use of MSC for the generation of ACL replacements will become the norm, perhaps using SF MSC since they would be in the intraarticular environment of the ACL to start with and thus, perhaps be positioned to participate in endogeneous repair already.
In a number of studies resulting from investigations from the Nakamura laboratory at Osaka University, and in collaboration with the author, Tissue Engineered Constructs (TEC) for cartilage, meniscal and bone repair have been generated from MSC derived from synovial membranes or synovial fluid [64-68]. The model of choice for these studies has been pigs, a species with a physiology similar to humans, and a large animal with knees also similar to humans, but similar TEC can be generated from human MSC [
While the above TEC approach is promising for repair of damaged/injured cartilage and menisci (and potentially tendons and ligaments as well), the implanted TEC would still have to overcome the challenges related to the inflammatory environment associated with the initial injury or the process of further degeneration following the injury (discussed above). In the pig model studies, such influences did not appear to be critical to the survival of the TEC, but it certainly could be a factor in possibly older patients, or those with more extensive damage where endogenous cells were being recruited to augment the functionality of the implant.
A number of cellular and tissue engineering approaches have been investigated to facilitate repair of damaged/ injured tissues in the knee. Some of these, such as in vitro tissue engineering for repair of cartilage and menisci are quite promising based on current findings. Other approaches using MSC may be effective in modifying symptoms associated with knee injuries or degenerative disease processes in the knee, but the possibility that such approaches will lead to structural repair, particularly in the absence of an appropriate scaffold, is not great in the current understanding of how the joint functions. One possibility to enhance the functioning of autologous MSC to facilitate repair may be to generate an acellular scaffold via tissue printing of appropriate dimensions (derived from imaging modalities), and possibly containing growth factors or homing molecules, which could be implanted and then autologous MSC injected to populate the implanted template. Thus, the appropriate template scaffold could be generated separate from the cells, and then repopulated with undifferentiated or in vitro differentiated MSC. The undifferentiated cells could then differentiate under more in vivo conditions.
Due to the complexity of the in vivo environment and the processes involved in tissue repair, it is not likely that anti-sense or limited target specificity approaches will have much benefit to repair the tissues of the knee. However, a combination of tissue engineering approaches + MSC, and with the control of the inflammatory environment within the knee may offer the best hope for the future of knee repair with a return to optimal functioning of this organ system [1,16].
The author thanks the many colleagues at the University of Calgary and elsewhere (particularly at Osaka University and at the Technical University of Dresden), trainees and laboratory personnel who have contributed to the studies associated with his laboratory over several years. He also thanks the Canadian Arthritis Network, The Arthritis Society, Canadian Institutes of Health Research, and Alberta Innovates Health Solutions for their financial support in performing studies at the University of Calgary. The author also apologizes to those investigators who have made significant contributions to the fields discussed, but were not cited due to space limitations.