American Society for Bone and Mineral Research-Orthopaedic Research Society Joint Task Force Report on Cell-Based Therapies

Task Force charges

The objective of this Task Force report is to provide guidance to investigators, clinicians, and the general public about the potential and challenges of cell-based therapies for both soft and mineralized skeletal structures. See Table 1 for the complete Task Force charges that informed development of this report and the recommendations included herein. Candidate sources of stem/progenitor cells are reviewed, and optimized experimental protocols for assessing their progenitor and healing properties in animal models are presented. The Task Force examined the diverse range of affected tissues to assess the current state of cell therapy and developed recommendations for investigators and observers regarding evidence of potential clinical efficacy of cell based therapy

Task Force review process

Overview of Task Force findings and recommendations

Task Force recommendations are summarized in Table 3. Guidance for conducting preclinical and clinical studies of cell-based therapies is offered as a part of these recommendations (see Key Question 5, below). The Appendices A through F provide more details regarding publications related to cell based therapies in various musculoskeletal tissues and also detail supplementary information on how to determine the cell autonomous and non-autonomous effects of a donor population being used for bone regeneration.

Conclusions

The efforts of this Task Force are designed to provide researchers and clinicians with a better understanding of the current state of the science and research needed to advance the study and use of cell-based therapies for skeletal tissues. The design and implementation of rigorous, thorough protocols will be critical to leveraging these innovative treatments and optimizing clinical and functional patient outcomes. In addition to providing specific recommendations and ethical considerations for preclinical and clinical investigations, this report concludes with an outline to address knowledge gaps in how to determine the cell autonomous and non-autonomous effects of a donor population used for bone regeneration. Currently, there is no proof of efficacy of these treatments given the lack of rigorous clinical studies and randomized clinical trials, and that these therapies should thus be considered experimental.

Stem/progenitor cells

Based on the fact that virtually all tissues undergo renewal, albeit at highly variable rates, enduring tissue regeneration depends on the presence of a subset of tissue-specific stem/progenitor cells within a given population that can fuel tissue renewal. Hence, considerable effort has been focused on identifying cell sources that have the ability to maintain tissue homeostasis, including musculoskeletal tissues.4 Stem/progenitor cells in mammals can be divided into two broad categories: pluripotent stem/progenitor cells and postnatal stem/progenitor cells from various sources (see https://stemcells.nih.gov/info/basics/1.htm for more information).

Pluripotent stem cells (embryonic stem cells) and induced pluripotent stem cells

Embryonic stem cells (ESCs) are pluripotent stem cells derived experimentally by extraction from the inner cell mass of an early-stage embryo, the blastocyst, whereas postnatal stem/progenitor cells are found in different organs and tissues. Pluripotency refers to the potential of a stem cell to differentiate into cells of all three germ layers—endoderm, mesoderm, and ectoderm. In addition to ESCs, induced pluripotent stem cells (iPSCs) can be generated from adult somatic cells by reprogramming with essential pluripotency transcription factors by a variety of methods and are nearly identical to ESCs.

Because of their highly uncommitted state, pluripotent ESCs and iPSCs present more challenges in terms of being induced to differentiate into a specific musculoskeletal cell type compared to multipotent stem/progenitor cells, such as postnatal stem/progenitor cells. Osteogenic differentiation provides an example. There have been a number of reports on the differentiation of human ESCs (hESCs) and iPSCs into osteogenic cells through the formation of embryoid bodies, spontaneous differentiation, indirect co-culture with osteogenic cells, treatment with conditioned medium generated by osteogenic cells, or use of various schemes for direct osteogenic differentiation.5 However, for the most part, the results have relied on in vitro differentiation assays that may not reflect capabilities in vivo, and results from limited studies involving in vivo transplantation are not conclusive.6-9 Furthermore, most studies use ectopic transplantation sites rather than sites within an injured skeletal tissue that would better mimic a clinical scenario. The lack of differentiation into a functional tissue is likely because these cells have not undergone a developmental process that commits them to a particular lineage.10 Thus, there is the issue of differentiation efficacy. Another concern is safety, such as (i) the potential for tumor formation, and (ii) the potential for immunological rejection when allogeneic ESCs or iPSCs are introduced in vivo.11 More standardization and safer methods are clearly needed to bring pluripotent stem cells into clinical use.

Connective tissue–specific stem/progenitor cells

BMSCs are non-hematopoietic adherent cells first identified and characterized by Friedenstein12 A subset of BMSCs are skeletal progenitor cells that differentiate into cartilage, bone, hematopoiesis-supportive stroma, and marrow adipocytes based on rigorous clonal and differentiation assays performed in vivo (bone, fat) and in vitro (cartilage). However, the original concept of a tissue-specific stem/progenitor cell for bone, was later altered to include other mesodermal derivatives such as muscle, tendon, and ligament, and BMSC terminology was altered so that these cells are considered synonymous with MSCs. A better justification and nomenclature of stem/progenitor cells from other connective tissue sources need to be developed to clearly differentiate the source of a progenitor population used for a particular cell therapy.

A number of markers of BMSCs have been identified (such as CD73, CD90, CD105, and CD146, along with lack of expression of CD11b, CD14, CD19, CD34, CD45, and HLA-DR).13 However, the markers used are neither specific nor exclusive, because they are also expressed in many adherent fibroblastic cell populations. Although such “markers” have been utilized to identify cells derived from periosteum, synovium, dental pulp, and periodontal ligament cells (tissues associated with the skeleton) that are very similar to BMSCs and are capable of differentiating into cartilage, bone, and fat in vitro,14 it should be noted that these various cell types are not identical in their in vivo differentiation capacity.15-18 Furthermore, the standard in vitro assays used to claim “tri-lineage” differentiation are often not predictive of in vivo differentiation capacity.19 Whether and how these differences may have arisen from their different native tissue microenvironment needs to be examined.

In the last several years, lineage tracing studies have identified a self-renewing multipotent cell population called the skeletal stem cell (SCC) present in bone tissue of humans (PDPN+, CD146–, CD73+, and CD164+) and mice (Tie2–, integrin AlphaV+, Thy–, 6C3–, CD105–, and CD200+) that has differentiation restricted to the osteoblast, chondrocyte, and stromal cell lineage, and not adipocytes. In both mice and humans, these cells, when implanted into the renal capsule form ectopic bone and cartilage, and the cells support marrow formation. The SCC population is expanded in culture by BMP-2 and in vivo by bone fracture.20-22 Another study identified a rare population of Gremlin 1+ (Grem1+) cells in the metaphysis and tissues adjacent to the growth plate in mice. Similar to the SCC population, Grem1+ cells are BMP-2 responsive, self-renewing, and formed bone, cartilage, and stromal tissues, with limited adipocyte differentiation. During development and aging, the Grem1+ population generates articular and growth plate cartilage, and is found in periosteum, osteoblasts, and osteocytes. Like SCC, the Grem1+ cells expand in fractures. Moreover, Grem1+ cells transplanted into fractures engraft, self-renew, and form osteoblasts, and Grem1+ cells can be harvested and re-expanded. The elimination of Grem1+ cells in developing mice results in reduced bone mass.23

Thus, there is a rare self-renewing SCC population in bone that is separate from the bone sinusoids that does not strictly meet the definition of an MSC, because it lacks capacity for adipocyte differentiation The SCC is necessary for development of bone and cartilage, maintenance of the adult skeleton, and bone repair. These approaches show the utility of cell lineage tracing in defining progenitor cell populations in vivo. However, the manner in which these rare cell populations are applicable for a cell-based therapy approach remains to be determined. Similarly, it is unclear whether an analogous tissue lineage–restricted stem cell is present in other musculoskeletal tissues, such as tendon, ligament, or disc.

Nonskeletal tissue–derived MSCs

MSC-like cells obtained from nonskeletal sources (eg, muscle, cord blood, Wharton's jelly, dermal, adipose, and amniotic fluid MSCs) have also been promoted as another source of progenitors for cell therapy of skeletal tissues.24, 25 However, many of the cells used in these studies were pretreated with chondrogenic/osteogenic factors, and whether these nonskeletal cells produce functional bone and associated tissues, or whether they induce local cells to undergo a repair process in vivo,26, 27 is not known.

Perivascular stromal cells from various tissues have stem cell characteristics and differentiate into osteoblasts, chondrocytes, and adipocytes. In bone marrow, cell lineage studies show that LepR+ perivascular cells located around both arterioles and sinusoids account for nearly all of the fibroblastic colony-forming unit (CFU-F) in bone marrow. Lineage tracing shows that LepR+ perivascular cells are the major source of bone and adipocytes in bone marrow in adult mice. Moreover, they are involved in the regeneration of bone marrow following radiation and participate in fracture healing.28 Perivascular cells isolated from fat, muscle, pancreas, skin, lung, brain, eye, gut, bone marrow, and umbilical cord are NG2+, CD146+, PDGF-R beta+, and alpha SMA+.29 These cells are multipotent and differentiate into osteoblasts, chondrocytes, and adipocytes. In addition, a Gli1+ stromal cell population present in the adventitial layer of arteries has stem cell characteristics. Gli1+ cells differentiate into osteoblasts and are responsible for the vascular calcifications that occur in atherosclerosis and in chronic kidney disease.30 Thus, perivascular stromal cells, as well as adventitial cells, are a key source of MSCs in nonskeletal tissue–derived tissues. These cells have been used in animal models in cell-based therapy approaches.31-35

Differentiated skeletal cells

Fully committed cells associated with the skeleton, such as osteoblasts, chondrocytes, and tenocytes, have limited ability to self-renew. However, they may be considered therapeutically useful in situations where there is low tissue turnover. ACI, which uses culture-expanded autologous chondrocytes harvested from the less weight-bearing articular cartilage of the joint, has shown efficacy in the repair of focal cartilage defects.3 It is, however, generally acknowledged that the cultured chondrocytes undergo dedifferentiation and/or hypertrophy in vitro, thus compromising the quality of the regenerate cartilage, which is often more fibrous in nature or sometimes undergoes overgrowth and/or calcification. Also, currently ACI procedures are recommended only in young patients and exclude older adults (≥45 years old),36-38 because this age group has a higher susceptibility to degenerative joint diseases.

Platelet-rich plasma

Derived from megakaryocytes, platelets contain a long list of (approximately 300) growth factors, such as platelet-derived growth factor, transforming growth factor beta, and vascular endothelial growth factor, and others such as cytokines, coagulation factors, fibrinolytic factors and proteins, proteases, antiproteinases, and lipids. Upon platelet activation, eg, as a result of tissue injury, these factors are released for acute repair but are unlikely to have sustained activity or localization. Consequently, there are efforts to develop preparations, such as platelet-rich plasma (PRP) gels or PRP combined with other biological materials such as demineralized bone matrix or cells, to extend the biological activity of PRP.43 PRP has been used in a number of clinical settings for bone regeneration44 cartilage repair,45 treatment of osteoarthritis,46 and tendon/ligament and meniscal repair.47 However, the biological action of PRP, eg, in regulating differentiation of multipotent MSCs,48 remains unclear. In addition, the efficacy of PRP in the repair of hard tissues has yet to be established, in view of the wide variability in methods and measures of the studies that have been conducted.

Conditioned medium

Cells from connective tissues (including skeletal tissues) secrete a broad array of growth factors, cytokines, and other biologically active factors. It has been suggested that conditioned medium can exert a beneficial effect on healing of skeletal tissues based on paracrine, immunomodulatory, and immunosuppressive effects that encourage local cells to begin the repair process. A large number of animal studies49 and one human study50 have used this approach to treat injured skeletal tissues; however, efficacy of this type of therapy remains unestablished.

EVs

Perhaps the first description of EVs (eg, exosomes, ectosomes, microvesicles, microparticles) in the literature comes from the finding of matrix vesicles in cartilage.51 EVs, first termed microparticles, have since been detected in many bodily fluids. It is now thought that in culture, virtually all cell types release some sort of EV, formed by different processes and of varying size. Results derived from the use of conditioned medium may be due to the presence of EVs in addition to other factors. EVs are proposed to mediate intercellular communications and have been implicated in repair and disease processes by virtue of the cargo that they carry, which may include proteins, biofactors, and RNAs, including specific miRNAs, among others.52, 53

Bone

Mice are commonly used to study cell-based therapies (eg, BMSCs, adipose-derived stromal cells) in models of osteoporosis, osteogenesis imperfecta, fracture healing, and ectopic bone formation (subcutaneous or intramuscular). Mice allow genetic approaches and have a high tolerance to xenograft engraftment. Humanized mice permit modeling of human cell activity and function. The reviewers recommend that the animal model yield convincing evidence with regard to bone architecture and mechanical properties. Impactful studies should clearly define the changes associated with cell therapy and the mechanism of cell involvement in the regenerative process.

A disadvantage of mice is that confirmation of findings typically requires subsequent experimentation in a larger animal model which more closely mimics the structure and mechanical features of human tissues prior to consideration for use in clinical trials. The rat is also an inexpensive small-animal model, and because it is relatively larger, it may have some advantages for mechanical testing. Gene-editing technologies are now being applied to the rat to provide genetic models that previously were only available in the mouse. Successful outcome should then be tested in larger animal models.

The models of bone disease/injury identified in the review included calvarial defects (46 studies), fracture (seven studies), heterotopic ossification (intramuscular, three studies), heterotopic ossification (subcutaneous, 38 studies), long-bone cortical defect (drill, 13 studies), long-bone segment (acute repair, 28 studies), long-bone segment (chronic repair, three studies), osteoporosis (nine studies), or other (heterotopic, two studies) (Appendix B).

Cartilage

Rabbits and rats were the most commonly used animals among the 49 preclinical studies that examined cell-based therapies for the regeneration of articular cartilage (Table 4). Rabbit models were the most frequently used. Rabbits are widely available, are relatively low-cost, have simple handling requirements, and have a robust base of literature for comparison.56 However, there are several disadvantages inherent in rabbit studies. Pure chondral lesions are difficult to create in the thin cartilage of this species. Also, researchers need to consider the potential for a natural healing response given the smaller size of cartilage defects (<3 mm)3 and the difference in mechanical loading in the rabbit knee relative to humans. Although rodents are widely used to screen new biomaterials and cell-laden constructs, the use of rodents for articular implants is less practical because of their very thin articular cartilage.57, 58 Furthermore, unlike larger animals, rodent growth plates remain open during adulthood. The epiphysis is therefore more highly vascularized, a feature that may contribute to more robust intrinsic cartilage repair.

Considering clinical relevance, a larger animal model is preferred to approximate the area and thickness of articular cartilage in the human.59 Full-thickness cartilage defects can be created without damage to the subchondral plate in dogs.60 However, the dog is rarely used for cartilage repair61 because its status as a companion animal makes it a less attractive option. Goats and sheep are frequently used in cartilage repair models because the knee joints are large enough to create lesions as large as those treated in patients. The larger defect size and thicker cartilage layer permit biochemical assays of cartilage repair tissue as well as biomechanical testing.62

Domestic pigs, minipigs, goats, and sheep, although more expensive, are attractive because of the thicker articular cartilage that more closely mimics the human joint.63 The main advantages of the horse model are the large joint size and thick articular cartilage layer with easy arthroscopic joint access, as well as the actual clinical need in equine veterinary care.64, 65 Horses can be monitored with respect to the clinical pain response to cartilage repair. A second-look arthroscopy with biopsy is usually the cornerstone of equine studies, allowing assessment of repair progression. However, specialized facilities and care are usually required, and thus this model is often used for late-stage development and pivotal studies.

In our review of 49 preclinical studies of cell-based therapy to promote cartilage repair, two used horses, six used sheep, 15 used rabbits, 10 used rats, six used goats, eight used pigs, one used dogs, and two used mice (Appendix C). Most of the models used an osteochondral defect model (42 studies). Five other studies used a partial-thickness cartilage-defect model, whereas two others used a model of global osteoarthritis from either injury (anterior cruciate ligament [ACL] and medial meniscectomy) or from chemically-induced osteoarthritis.

A total of 14 human studies were identified that examined cell-based therapy for articular cartilage repair. Because ACI therapy is already an FDA-approved procedure, studies using articular chondrocytes as a cell source were not included. All of the studies included full-thickness osteochondral defects. Bone marrow aspirate was the most common source of cells (10 studies) followed by peripheral blood–derived progenitors. One study each used either nasal chondrocytes or synovial progenitor cells. In 6 studies, mesenchymal progenitor cells from bone marrow (four studies), synovium (one study), and nasal cartilage (one study) were expanded in cell cultures prior to delivery into the joint.

Each of the papers was scored according to whether the study assessed the following features in the context of cartilage repair: radiographic criteria, histology, origin of the cell-tissue organization, biochemical criteria, functional outcomes, fate of the transplanted cells, and non-target tissue effects. Each criterion was scored as 0 (not examined); 1 (superficial analysis); or 2 (detailed analysis). A range of scores was identified, but consistently show higher scores in the preclinical models. All of the human studies had level of evidence 3 or 4, and analysis was limited because of the morbidity associated with tissue procurement in humans.

Intervertebral disc

Although there are murine genetic models of intervertebral disc disease, they often lack normal disc formation or develop severe structural alterations during development and thus have not been extensively used to study the role of cell-based therapies for disc regeneration.66 The small size of mouse and rat discs compared to human discs limits their use in the assessment of disc repair methodologies, because nutrient diffusion is not fully evaluated. In addition, the morphology of the rodent disc differs somewhat from humans. Although rabbits and guinea pigs are larger, their discs are still several-fold smaller than human discs, and the tissues do not experience the same degree of loading. In comparison, larger, quadruped animals such as sheep have intravertebral discs that are closer in size and loading to those of humans.

Cell-based therapy for disc repair has been reported in at least 50 studies. Animals utilized include sheep (one study), dogs (five studies), rabbits (19 studies), minipigs (four studies), rats (10 studies), and mice (seven studies) (Appendix D). Models of disease occurring through either natural means (ie, aging as in human disc degeneration) or induced by removing various amounts of nucleus pulposus (NP) tissue (nucleotomy; the majority of animal studies), removal of herniated tissue (at the time of discectomy, two human studies), stab injuries to the annulus fibrosis with or without removal of NP tissue (two studies), or removal of a whole disc (three studies). Most studies tested treatment in the acute injury setting, but in three studies, cell therapy was used at 4, 6, or 12 weeks postinjury in dogs.67-69 Chondrodystrophic canines have been used, but whether these animals had developed premature degenerative disc disease at the time of implant evaluation was not specified.

Meniscus

Rabbits (43%) and rodents (16%) have been frequently used in studies that evaluate the underlying biological and molecular mechanisms of meniscus healing. Rabbits have also been used for preliminary analyses of new stem-cell or progenitor-cell therapy and tissue engineering approaches for the treatment of meniscal injury. Rabbits are less expensive than larger animal models and allow for the use of the larger group sample sizes that are necessary for more comprehensive outcome measures including histology, biochemical and molecular analyses of the repair tissue, and biomechanical testing.

Sheep (8%), pigs (8%), and primates (4%) have been used as large animal models to study meniscus healing. These animals possess knee joints similar in size to the human knee, enabling testing of implants that could be used in humans. However, unlike humans, these animals are quadrupeds with obvious differences in meniscus and cartilage contact mechanics when compared to humans. There is also an inability to control postoperative weight-bearing in most large animal models.

For meniscus, the majority of models used were local acute injury (19 studies). Chronic injury studies (four studies) were also examined. Most meniscus models evaluate tissue formation in a punch defect, although linear tear models have also been used (Appendix E).

Tendon/ligament

Smaller animal models, including rabbits70-81 and rodents (rats and mice),12, 82-97 have been used to evaluate the effectiveness of stem-cell/progenitor-cell therapies in acute tendon injury and repair. Early-stage studies and anatomical considerations might necessitate the use of smaller animals, such as in the case of rotator cuff repair, where the rat is a commonly used model because it resembles the human shoulder anatomy.98 Furthermore, if the objective is to track the fate of implanted cells and to mechanistically evaluate their contributions to the repair tissue, rodent models (rats and especially mice) have the distinct advantage of the availability of reporter gene models (eg, GFP, RFP, mTmG, nTnG) and genetic models of conditional gene deletion.

Large animal models, including equine,99-102 ovine,100, 103 swine,104, 105 and canine,106, 107 might better meet the FDA's preference for demonstrating efficacy of cellular therapies and delivery devices prior to human approval. The suitability of an animal model for tendon/ligament repair depends on the tissue being examined and the objectives of the study. For example, the canine model is commonly used to study flexor tendon repair because of the ability to simulate zone II injuries, reproduce clinical protocols of multistrand suture repair, and implement physical therapy.106, 107 Similar arguments can be made in favor of using large animals (eg, pigs) for evaluating cellular therapies in ACL repair and reconstruction.104

For tendon/ligament, various acute injury and repair models have been examined, including flexor tendon,72, 73, 99-101, 105-107 rotator cuff tendon,83-86 Achilles tendon,12, 75, 81, 89, 91, 95, 103 patellar tendon,70, 71, 74, 76-80, 87, 88, 94, 96, 97 and ACL.92, 93, 104 Furthermore, a number of equine studies have focused on veterinary clinical applications of autologous stem/progenitor cells to treat chronic overuse injury of the superficial digital flexor tendon in racehorses using ultrasound-guided intratendinous injections99-102 (Appendix F).

Recommendations for preclinical studies

Recommendations for clinical studies

Research and clinical ethical considerations

Preclinical animal models

Overall, the reviewers found no preferred, standard animal model for a preclinical study. Thus, the Task Force recommends that, in addition to size and anatomical considerations, the choice of the animal model should thoughtfully consider other aspects of the study objectives and design, including the cost, technical challenges, the potential to use both autologous and allogeneic cells, and the degree to which the model mimics human disease.

An optimal experimental approach to evaluating stem/progenitor cells for enhancing musculoskeletal tissue repair/regeneration would be to initiate studies in small animals that focus on cellular, molecular, functional, and mechanical outcome measures, and allow the examination of genetic factors influencing regeneration. Once these models provide proof of principle for the utility of a specific cell population in augmentation of repair, additional investigation would be completed in larger animals that more closely model the anatomical size and weight-bearing characteristics of human skeletal tissues. Subsequent successful outcomes in large-animal models, with inclusion of appropriate safety and efficacy profiles, would identify prime candidates for human clinical trials.

In studies where the reparative potential of human cells is assessed, immune reaction in the animal model is a challenge. Human cells are sometimes assessed in less-optimal models, including immunodeficient nude rat models or larger animals with drug-mediated immunosuppression, conditions that may influence the repair process. Still, with regard to immunological considerations of the efficacy of allogeneic progenitor cell therapy, mouse models of selective depletion of a subpopulation of T cells108 or pig models of genetically manipulated major histocompatibility complex (ie, swine leukocyte antigen)109 can be particularly useful.

The FDA will require clear and unequivocal evidence for the cellular basis of stem/progenitor–based therapies for musculoskeletal tissue repair. A combination of models will provide support for cell-based therapeutics, including in vitro three-dimensional models of tissue regeneration/tissue chip models,110 small-animal models; and large-animal models. Fundamental requirements for adequate interpretation of cell-based therapies include the ability to: (i) track the donor cell population; (ii) determine cellular fate and differentiation; and (iii) define the role of the cell population in restoring a stable and mechanically functional tissue. (Additional information about determining cell fate can be found in Appendix A).

Interpretation of the role of donor stem/progenitor cells in the repair process is critical. Persistence of donor-derived cells during the repair should preferably be observed, although non–cell autonomous mediation of therapeutic effects must also be considered. Donor cells may undergo terminal differentiation and remain in the tissue as mature cells or may have a critical early and more transient role in driving the repair process. Cell-based therapy should be investigated in relevant models across the various mesenchymal tissues. Because mesenchymal tissues (ie, bone, cartilage, tendon, ligament, and disc) each require unique mechanical properties, the model should permit assessment of whether an appropriately functioning tissue forms. There is a need for the development of additional in vitro and in vivo models, as well as computational approaches. In particular, noninvasive assessments of tissue composition, structure, and function need further development as specific cell-based therapies are extended to human trials.

Large-animal models, although essential, are expensive and technically demanding. Large-animal studies are currently limited to autologous-based experiments, in which markers distinguishing host from donor are limited. The NIH-supported National Swine Resource and Research Center, a large-animal resource, has been developed. The center can provide pig lines with GFP reporters and also has pigs with genetic backgrounds similar to the mouse NSG.NOG strain. Successful results from studies using human progenitor cells in the mouse could transition to the immunocompromised pig as a route to FDA-approved clinical trials of a cell-based therapy.

Human clinical trials

Appropriate assessment of cell-based therapies in human studies will require appropriate study design, including the use of appropriate controls, blinding and elimination of investigator bias, appropriate validated outcome measures, and rigorous statistical analysis. Because of the difficulty and/or morbidity associated with harvesting tissues from human subjects, the development and use of sensitive, noninvasive measures of tissue composition, structure, and mechanical function is essential.

In humans, sensitive, validated PROMs will provide a powerful tool to assess functional improvement. PROMs utilize patient-based assessments of their own functional status and perception of pain. Many validated outcome measures are now commonly used to assess the efficacy of various medical and surgical therapies. For upper extremities, functional outcome measures include grip strength and assessment of arm, shoulder, and hand function (ie, Disabilities of the Arm, Shoulder, and Hand [DASH] score) based on questionnaires. Lower-extremity fracture functional assessments could include the Timed Up and Go Test or the Lower Extremity Functional Scale. The Harris Hip Score and the 36-item Short Form Survey (SF-36) can measure hip and lower-extremity PROMs of function. Finally, functional criteria to assess the success of a spinal fusion could include the Oswestry Disability Index or the SF-12 or SF-36. Through a Common Fund initiative, the NIH developed a comprehensive PROM system called PROMIS that uses computer-adapted testing and thus has increased efficiency and sensitivity. This is being increasingly adopted as a PROM for patients with musculoskeletal disease.

The primary clinical goals of cell therapies are to treat symptoms, especially pain and instability, and to allow for return of normal function of the targeted tissue. Success can be defined as: (i) improvements in pain and/or physical function; (ii) a durable response (years); and (iii) a return of structural support or joint mobility that permits movement. Structural repair is a secondary outcome, implying that the tissue repair or implanted tissue has value to the degree which it permits painless function. The individual characteristics and properties necessary for regeneration are specific to the various musculoskeletal tissues (Table 4).

For therapeutic product development, strong preclinical evidence in model organisms on healing and function is an absolute requirement for the initiation of randomized clinical trials, which involve objective clinical outcome and follow-up measurements. Rigorous, scientifically based understanding of the mechanism of action is ultimately required to justify the therapeutic application of cell-based therapy. It is noteworthy that adult stem/progenitor cell therapy, which is often being practiced in an unregulated, non-standardized manner, has recently resulted in several highly publicized studies. To ensure safety and enhance efficacy, adult stem-cell therapy should be practiced only in a regulated and standardized manner. In addition, clinical trials submitted for publication should adhere to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) or Consolidated Standards of Reporting Trials (CONSORT) guidelines1.

Ethical considerations