All musculoskeletal tissues contain living cells situated within a dense ECM. The phenotype of these cells is, by definition, appropriate for the generation of the ECM that defines the specialized function of the host tissue—this is their primary role. Given this, these are the most ideal source of regenerative cells after injury. Indeed, in several current clinical applications, endogenous cells are the therapeutic cell type of choice. These primary cells are removed from the native tissue, expanded in culture, coupled with a delivery system (in some cases), and then applied to the defect site.¹ For example, several FDA-approved cartilage repair procedures rely on the isolation of primary cells, wherein morsels of cartilage from a non-load-bearing region of the affected joint are removed (using an arthroscopic procedure), followed by extraction of living chondrocytes from these native tissue segments, and expansion in an in vitro setting. These cells are then either delivered directly or are seeded onto a provisional delivery
material and then reimplanted into a focal defect in the same joint. The recent approval of the matrix-assisted autologous chondrocyte implantation (MACI) procedure represents the first “tissue engineered” produce in clinical use in the musculoskeletal system.²

While primary cell-based therapeutics are already impacting clinical practice, several limitations do exist in this approach. Unlike most tissues in the body, the density of primary cells within adult musculoskeletal tissues is quite low, on the order of 10% of the volume or less.³ This is due to both the primarily mechanical roles of these tissues as well as the relatively limited anabolic requirements of an already established musculoskeletal tissue. Once growth and development is complete, the role of endogenous cells is to replace the ECM in localized areas to address tissue damage. This occurs at a slow rate, however, and wholesale replacement of the ECM on a regular basis does not occur. Recent studies using carbon isotope analysis, taking advantage of the above-ground nuclear testing of the 1950s and 1960s, showed that the majority of the collagen in soft tissues that is present at the time of skeletal maturity remains present for the rest of life.⁴ This suggests that, in healthy tissues in adults, cells operate at a low level of activity, and expansion in cell number is rare.

As there are so few cells, and those that are present operate mostly to maintain homeostasis, the application of these cells to in a regenerative context raises some issues. First, to populate large defects, primary cells must be dramatically expanded in number. The conditions under which this expansion occurs can impact the ultimate therapeutic efficacy of these primary cells. When primary cells undergo a high number of population doublings in vitro, typically on rigid tissue culture surfaces, cellular exhaustion (senescence) rapidly emerges. This results in the cessation of replicative capacity and the loss of cellular metabolic function.⁵ Moreover, when primary cells are isolated from individuals of an advanced age, this in vitro senescence emerges more rapidly or may limit the number of population doublings that these cells can undergo. Even if expansion does not instigate senescence, the phenotype of primary cells is often lost through the expansion process. This is particularly true for chondrocytes that, within a few cell divisions, undergo a reversion to a fibroblast-like phenotype and cease production of proteoglycans and type II collagen (two essential components of the cartilage ECM).⁶ Given that these cells must, de novo, generate new ECM at a rapid rate post implantation, this may compromise in vivo function of the regenerate tissue.

The use of primary cells has already impacted orthopaedic clinical practice. Current advances are focusing on overcoming the limitations that currently exist to make the process more efficient and cost-effective and to widen the indications for these cells. For example, work is being done to define the most optimal set of biological factors (growth factors, cytokines, etc.) and culture environments (material framework, see below) that can improve expansion efficiency while preserving phenotype. The delivery systems for these primary cells has also become more sophisticated, with newer autologous treatments for cartilage using biomaterials as a carrier to improve targeting of primary cells to the defect site and to decrease surgical complexity.⁷ Likewise, advances in drug delivery have made it possible for these primary cells to be delivered along with agents that release factors to increase retention and anabolic activity post implantation. Others have explored coculture systems where, for example, primary chondrocytes are mixed with progenitor cells (described below) to take advantage of the need for fewer cells at the time of implantation,⁸ allowing for these to become
one-stage procedures.⁹ Finally, with this and other approaches, methods are being developed to select the most efficient cartilage forming cells, including from alternative sources (such as nasal cartilage),¹⁰ as well as to remove from the population those showing senescent or other aberrant phenotypes.¹¹

While primary cells, by definition, possess the appropriate phenotype for tissue repair, in many instances the isolation of these cells is not feasible, or the isolation and expansion process may not be compatible with retention of phenotype and clinical application. To overcome this limitation, work over the last two decades has focused on the identification and application of stem cell or stem cell-like populations for musculoskeletal tissue repair. A stem cell is a cell that can replicate indefinitely to produce progeny that both maintain the original stem cell pool and/or can be induced to differentiate into a phenotype resembling the primary cells in the target tissue of interest. When the spectrum of fates is limited, these cells are called multipotent, to indicate the limited set of phenotypes that can be achieved, and are referred to most appropriately as progenitor cells (rather than stem cells).

Most popular among these cells are bone marrow-derived mesenchymal stem/progenitor cells, colloquially termed MSCs. These cells, first identified in the 1960s and 1970s,^12,13 can undergo lineage specification into cells resembling bone, fat, and cartilage cells, among others.^14,15 Lineage specification of these cells is generally directed using specific soluble factors in combination with culture environments permissive of a given lineage—for instance, the cartilage phenotype is best promoted when these cells are cultured in high-density micromass or pellet cultures or in a suitable three dimensional (3D) environment that can promote the rounded cell shape of a chondrocyte.¹⁶ With appropriate induction, these progenitor cells can generate a cartilage-like matrix that increases in mechanical properties over time, resembling native tissue in terms of composition, mechanics, and molecular profile.^17,18

Although the bone marrow is by far the most common location for sourcing of these progenitor cells, this is not the only location in the body from which multipotent cells can originate. In the early 2000s^19,20 it was found that lipoaspirate contained a population of mesenchymal-like progenitors as well, with similar differentiation capacity as those isolated from bone marrow. Since then, nearly every tissue in the musculoskeletal system has been shown to harbor a population of so called “endogenous stem cells” that can differentiate into cartilage, fat, and bone and are speculated to play a role in endogenous repair of these native tissues. For this reason, these cells are widely used as a starting source for many repair applications.

Mesenchymal progenitors show great promise as a cellular therapeutic and for growing new tissues ex vivo for reimplantation. Given that bone marrow and adipose tissue scan be readily harvested in a relatively noninvasive manner, they also serve as a clinically relevant source of autologous starting cells. Moreover, clinical trials have shown the safety of joint injections of allogeneic MSCs in humans,²¹ with recent trials showing some evidence of efficacy in some patients, though comprehensive and well-regulated trials are limited.²²

However, limitations in this cell type do exist. First, these cells are by definition only multipotent with fixed lineage potential, not totipotent, and so can only differentiate along specific lineages and may not be able to address all tissues in the musculoskeletal system. Additionally, the extent to which they take on the appropriate primary cell phenotype is often limited, where incomplete commitment to the differentiated lineage being a common outcome. For example, if a cell becomes an “inefficient” chondrocyte, production of matrix will only occur under ideal situations, and so in vivo production and survival in the demanding environment of the joint may be compromised.²³ Also, because commitment may be incomplete, this raises the potential for subsequent transition in phenotype after implantation. For example, it has been well documented that MSCs that have undergone in vitro chondrogenesis can transition to an osteogenic phenotype after in vivo implantation.²⁴ This is compounded by the variation seen in these cell populations both within a single donor, and between individual donors. As every MSC population initiates from a number of different individual cells (that expand into “clones”) and each of these starting cells can have a different potential, the population is by definition heterogeneous.²⁵ Likewise, between donors, marked variability exists, based on age of donor and other systemic comorbidities that impact the number of resident stem cells. Finally, like primary cells, mesenchymal progenitors are not “true” stem cells and so undergo replicative senescence and loss of potentiality with extended culture expansion.²⁶

As with primary cells, a number of technological advances are being used to improve the therapeutic efficacy of mesenchymal progenitor cell populations. High-depth and single-cell transcriptomic analysis has uncovered patterns of expression across populations of progenitor cells that may improve selection criteria of cells with the “best” phenotype.^27,28 Likewise, material substrates have been introduced that can maintain the “stemness” of these progenitor cells during culture expansion, potentially improving their long-term function.²⁹ Others have used biomaterial cues (more below) and drug and gene delivery methods to help to control the phenotype of these cells, both during in vitro differentiation and after in vivo implantation. Also, as with primary cells, coculture of mesenchymal progenitors with a very small fraction of primary cells appears to improve phenotypic specialization by the progenitors as well as the retention of this specialized phenotype.⁸ Finally, as the cells are typically expanded in vitro prior to being delivered back into a repair environment, they can
be engineered to express a number of factors, including those that promote their own differentiation or factors that improve the host environment to increase retention/survivability after implantation.³⁰ This could improve their functional role, but it should be noted that any cell type that is more than “minimally manipulated” will face a more stringent review by the FDA prior to approval for clinical use.

The final category of cells now in development for musculoskeletal applications is pluripotent stem cells (PSCs). These cells can undergo lineage specification to any cell type and can reconstitute an entire organism when integrated into the germline of a developing embryo. As such, these cells represent the greatest potential for repair across the musculoskeletal system.³¹ Multiple sources of PSCs now exist. The original PSCs were derived from the inner cell mass of developing mouse and human embryos and were termed embryonic stem cells (ESCs). After expansion on a feeder layer, and with careful and sequential exposure to developmentally relevant morphogens, ESCs can differentiate along all relevant lineages of the musculoskeletal system. Indeed, there are now multiple published protocols for pushing ESCs toward a cartilage-like phenotype.^{32,33,34,35,36,37,38} In addition to ESCs, the discovery of somatic cell “reprogramming” has enabled production of PSCs from nearly any cell in the body in adult sources. These induced pluripotent stem cells (or iPSCs) are generated (typically from skin fibroblasts) by transfecting with four critical transcription factors (the so called “Yamanaka factors,” based on the originator of this process).³⁹ This reprogramming step resets the epigenome of these cells (marks and histone organization that helps to instill differentiation) to a more primitive state, enabling their subsequent differentiation into any cell type via carefully staged induction processes that are similar to those used to direct lineage specification in ESCs. Indeed, a host of recent studies have reported on the efficient production of chondrocyte-like cells from iPSCs and have begun to show the potential of these cells in a number of regenerative scenarios.^40,41

Limitations in the use of PSCs are consistent with (and perhaps more apparent than) those in multipotent progenitor cell sources. This includes population heterogeneity, variation in differentiation potential based on source, and stability of phenotype after differentiation. Because PSCs have a greater potential for differentiation, emergence of alternative lineages may be more difficult to control. In addition, few ESC lines are available for federally funded use, and the clinical relevance of these cells remains limited. iPSCs offer a potentially autologous alternative, though the processes for reprogramming, expansion, and induction of a stable phenotype present a number of challenges in terms of reduction to clinical practice, despite considerable research activity in this area.