Mesenchymal stem cell-based therapies in regenerative medicine: applications in rheumatology

Growing knowledge on the biology of mesenchymal stem cells (MSCs) has provided new insights into their potential clinical applications, particularly for rheumatologic disorders. Historically, their potential to differentiate into cells of the bone and cartilage lineages has led to a variety of experimental strategies to investigate whether MSCs can be used for tissue engineering approaches. Beyond this potential, MSCs also display immunosuppressive properties, which have prompted research on their capacity to suppress local inflammation and tissue damage in a variety of inflammatory autoimmune diseases and, in particular, in rheumatoid arthritis. Currently, an emerging field of research comes from the possibility that these cells, through their trophic/regenerative potential, may also influence the course of chronic degenerative disorders and prevent cartilage degradation in osteoarthritis. This review focuses on these advances, specifically on the biological properties of MSCs, including their immunoregulatory characteristics, differentiation capacity and trophic potential, as well as the relevance of MSC-based therapies for rheumatic diseases.


Introduction
For several years, mesenchymal stem cells (MSCs; also called mesenchymal stromal cells) have been largely studied and used as a new therapeutic tool for a number of clinical applications, in particular for the treatment of rheumatologic disorders. MSCs indeed have therapeutic potential for bone and joint diseases due to their multipotent diff erentiation abilities and the secretion of a variety of cytokines and growth factors that confer on them anti-fi brotic, anti-apoptotic, pro-angiogenic and immunosuppressive properties. Th ey are currently being tested in several clinical trials for such diverse appli cations as osteoarthritis, osteogenesis imperfecta, articular cartilage defects, osteonecrosis and bone fracture. Moreover, good manufacturing practices for the production of clinical-grade MSCs at high expansion rates without transformation are now well established [1]. Here, we review the present knowledge on the mecha nisms underlying the therapeutic properties of MSCs and their applications in animal models and clinics in the fi elds of bone and cartilage repair, chronic infl ammatory or degenerative disorders, as well as genetic diseases.

Defi nition of mesenchymal stem cells: location and characterization
MSCs were fi rst identifi ed in the bone marrow (BM) [2] but are now described to reside in connective tissues and notably in adipose tissue (AT) [3], placenta [4], umbilical cord [5], dental pulp [6], tendon [7], trabecular bone [8] and synovium [9]. It has also been suggested that MSCs could reside in virtually all post-natal organs and tissues [10]. BM and AT are, however, the two main sources of MSCs for cell therapy due to high expansion potential and reproducible isolation procedures. Historically, the fi rst characterized MSCs derived from BM remain the most intensively studied and are still the reference. ATderived MSCs (ASCs) are easier to isolate in high numbers. Nevertheless, while they display characteristics similar to BM-MSCs, their transcriptomic and proteomic profi les show specifi cities particular to the tissue origin [11]. MSCs have also been described to reside in a perivascular location and to express markers specifi c for pericytes [12,13]. However, in AT, ASCs are mainly located in the stroma around the adipocytes and only few of them have a perivascular location. Importantly, in the tissue, none or very few ASCs express pericyte markers, even those that are located around the vessels [14].
MSCs are defi ned according to three criteria proposed by the International Society for Cellular Th erapy [15].

Abstract
Growing knowledge on the biology of mesenchymal stem cells (MSCs) has provided new insights into their potential clinical applications, particularly for rheumatologic disorders. Historically, their potential to diff erentiate into cells of the bone and cartilage lineages has led to a variety of experimental strategies to investigate whether MSCs can be used for tissue engineering approaches. Beyond this potential, MSCs also display immunosuppressive properties, which have prompted research on their capacity to suppress local infl ammation and tissue damage in a variety of infl ammatory autoimmune diseases and, in particular, in rheumatoid arthritis. Currently, an emerging fi eld of research comes from the possibility that these cells, through their trophic/regenerative potential, may also infl uence the course of chronic degenerative disorders and prevent cartilage degradation in osteoarthritis. This review focuses on these advances, specifi cally on the biological properties of MSCs, including their immunoregulatory characteristics, diff erentiation capacity and trophic potential, as well as the relevance of MSC-based therapies for rheumatic diseases.
First, MSCs are characterized as a heterogeneous cell population that is isolated by its property of adherence to plastic. In culture, MSCs are able to develop as fi broblast colony forming-units. Second, MSCs are distinguished by their phenotype: MSCs express the cell surface markers CD73, CD90 and CD105 and are negative for CD11b, CD14, CD34, CD45 and human leukocyte antigen (HLA)-DR. More recently, the CD271 marker was used to isolate highly enriched BM-MSC populations [16]. Whereas BM-MSCs are negative for the CD34 marker, native ASCs can be isolated according to CD34 expression, although this rapidly disappears with cell proliferation in vitro [14,17]. Th e third criterion to defi ne MSCs, based on a functional standard, is their capacity to diff erentiate into at least three mesenchymal lineages, namely bone, fat and cartilage.

Functional properties of mesenchymal stem cells
Diff erentiation capacity and paracrine signaling are both properties relevant for therapeutic applications of MSCs. MSC diff erentiation contributes by regenerating damaged tissue, whereas MSC paracrine signaling regulates the cellular response to injury.

Diff erentiation properties
MSCs are an attractive source of cells for bone and cartilage engineering because of their osteogenic and chondro genic potential. Th eir diff erentiation capacity is generally shown in vitro using specifi c culture conditions but also in vivo in diff erent animal models [18]. Besides this trilineage potential, MSCs can also diff erentiate into myocytes [19], tendinocytes [20], cardiomyocytes [21], neuronal cells with neuron-like functions [22,23] and other cell types. Th e diff erentiation potential is dependent on environmental factors, such as growth factors, but also physical parameters, such as oxygen tension, shear and compressive forces, and elasticity of the extracellular three-dimensional environment.

Paracrine properties
MSCs release various soluble factors that infl uence the microenvironment by either modulating the host immune response or stimulating resident cells.
Th e immunomodulatory properties of MSCs, character ized by the capacity to inhibit the proliferation and function of all immune cells, have been largely described both in vitro and in vivo (reviewed in [24]). Immunomodulation requires the preliminary activation of MSCs by immune cells through the secretion of the proinfl ammatory cytokine IFN-γ, together with TNF-α, IL-1α or IL-1β [25,26]. Th e induction of MSC immunomodulation is principally mediated by soluble mediators. Among these, indoleamine 2,3-dioxygenase has been shown to be a major player in human MSCs but absent or poorly expressed in murine cells, while nitric oxide is expressed at low levels in human MSCs but at high levels in murine MSCs following IFN-γ stimulation [26]. Transforming growth factor (TGF)-β1, hepatocyte growth factor (HGF), heme oxygenase 1, IL 6, leukemia inhibitory factor, HLA G5, IL-10 and IL-1 receptor antago nist (IL-1RA) as well as prostaglandin E2 have been proposed as other mediators involved in MSCmediated immunomodulation (reviewed in [24]). MSCs suppress B-and T-cell proliferation and alter their function, inhibit the proliferation of activated natural killer cells, interfere with the generation of mature dendritic cells from monocytes or CD34 + progenitor cells, and induce an immature dendritic cell phenotype [27,28]. Finally, MSCs inhibit Th 17 cell diff erentiation and induce fully diff erentiated Th 17 cells to exert a T cell regulatory phenotype [29].
Although soluble mediators are the main actors in MSC immunosuppression, cell-cell interactions have been shown to be involved in this process. Recently, tolllike receptor (TLR) stimulation has been shown to modulate the action of MSCs on the immune system. Indeed, TLR4-primed MSCs, or MSC1, mostly elaborate pro-infl ammatory mediators, while TLR3-primed MSCs, or MSC2, express mostly immunosuppressive ones [30].
Th e trophic properties of MSCs are related to the tissue regeneration process through bioactive factors. Th ese factors may act directly, triggering intracellular mechanisms of injured cells, or indirectly, inducing secretion of functionally active mediators by neighboring cells. MSCs are capable of attenuating tissue injury, inhibiting fi brotic remodeling and apoptosis, promoting angiogenesis, stimu lating stem cell recruitment and proliferation, and reducing oxidative stress. As an example, in a hamster heart failure model, intramuscularly injected MSCs, or even more importantly MSC-conditioned medium, signifi cantly improve cardiac function. Improvement occurred via soluble mediators acting on proliferation and angiogenesis, resulting in higher numbers of myocytes and capillaries, and on apoptosis and fi brosis, which were signifi cantly reduced [31]. Th e prominent factors identifi ed in these processes were HGF and vascular endothelial growth factor (VEGF). Th e authors demonstrate the activation of the JAK-STAT3 axis in myocytes, which increases the expression of the target genes HGF and VEGF [32]. Activation of the STAT3 pathway is crucial since its inhibition by TLR4 activation inhibits MSCmediated cardioprotection [33]. Secretion of VEGF by MSCs also attenuates renal fi brosis through immune modulation and remodeling properties in diff erent models of kidney injury [34,35]. Th e other mediators that are important actors during tissue remodeling and fi brosis formation are matrix metalloproteinases (MMPs) and tissue inhibitors of MMP (TIMPs). MSC-secreted TIMPs are capable of playing important roles both under physiological conditions in their niche and in pathological situations [36,37].

Chemotactic properties
Injured tissues express specifi c receptors or ligands that are believed to trigger the mobilization of MSCs into the circulation, facilitating traffi cking, adhesion and infi ltration of MSCs to the damaged or pathological tissues, in a mechanism similar to the recruitment of leukocytes to sites of infl ammation. In the damaged tissues, MSCs are believed to secrete a broad spectrum of paracrine factors that participate in the regenerative microenvironment and regulate immune infi ltration [38]. Administration of MSCs, either systemically or locally, has been reported to contribute to tissue repair, suggesting the need to enhance the pool of endogenous MSCs with exogenously administered MSCs for effi cient repair. A better understanding of MSC traffi cking and homing mechanisms should help in designing novel therapeutic options to compensate for a defi ciency in the number or function of MSCs that may occur in injured tissues.

MSCs for bone and cartilage repair
Interest in using MSCs for tissue engineering has been validated in numerous pre-clinical models and is under evaluation in clinics. At least 16 clinical trials are recruiting for the therapeutic application of MSCs for cartilage defects, osteoporosis, bone fracture, or osteonecrosis. For successful tissue engineering approaches, implantation of MSCs will require the use of growth and diff erentiation factors that will allow the induction of the specifi c diff erentiation pathways and the maintenance of the bone or chondrocyte phenotype together with an appropriate scaff old to provide a three-dimensional environ ment. Defi ning the optimal combination of stem cells, growth factors and scaff olds is thus essential to provide functional bone and cartilage.
Bone engineering strategies are warranted in cases of large bone defects or non-union fractures, which remain a serious problem as the associated loss of function considerably impairs the quality of life of aff ected patients. A vast variety of bone graft substitutes is already commercially available or under intense pre-clinical investigation to evaluate their appropriateness to serve as biomaterials for tissue engineering strategies (reviewed in [39]). Briefl y, bone substitutes are assigned to the group of either inorganic (mostly calcium phosphate-or calcium sulphate-based materials, or bioactive glasses) or organic matrices (natural processed bone graft or synthetic polymers). Moreover, it has to be stressed that the success of bone graft substitutes needs a fun ctional vascular network to obtain high quality osseous tissue.
Enhanced vascularisation is generally achieved by the provision of angiogenic growth factors that have been shown to increase bone healing [40]. To date, corticocancellous bone grafts remain the most frequently used way of reconstructing large bone segments. Despite promis ing reports on the potential of bone engineering, particularly for oral and maxillofacial surgeries, these innovative therapeutic strategies have so far been too sporadic, and with low numbers of patients, to give interpretable results. Further eff orts are needed to state more precisely the indications in which tissue engineered constructs could replace conventional therapies and improve clinical outcome of patients.
After traumatic or pathological injury, the capacity of adult articular cartilage to regenerate is limited. Th e current proposed surgeries (microfracture, osteochondral auto-or allografts, or cell-based therapies using chondrocytes) may lead to fi brocartilage and not restore hyaline articular cartilage in the long term. Several kinds of combined scaff olds have been evaluated for cartilage engineering using MSCs (reviewed in [41]). More recently, micron-sized fi bers, produced by the electrospinning technique, were shown to provide a structure and proper ties comparable to the cartilage extracellular matrix and to enhance chondrogenesis [42]. Eff orts are being made to improve scaff olds by combining several biomaterials (poly(lactic-co-glycolic acid) sponge and fi brin gel) with an inducing factor (TGF-β1) with satisfactory results [43]. Recently, our group has shown that MSC-coated pharmacologically active micro carriers releasing TGF-β3 implanted in severe combined immu nodefi ciency (SCID) mice resulted in the formation of cartilage, suggesting that they could repre sent a promising injectable biomedical device for cartilage engineering [44]. An alternative way to avoid direct transplantation of MSCs for tissue engineering is to recruit endogenous progenitor cells. Indeed, the replace ment of the proximal condyle in a rabbit by a TGF-β3-infused bioscaff old resulted, 4 months later, in a scaff old fully covered with avascular hyaline cartilage in the articular surface. Th e scaff old was also integrated within the regenerated subchondral bone, suggesting that the regeneration was probably due to homing of endogenous cells [45]. Although much progress has been made in the manipulation of cells and constructs for cartilage engineering, the generation of functional repaired tissue remains to be optimized.

MSCs for treatment of genetic diseases
Recent advances in stem cell research have prompted the development of cell-based therapies to replace cells that are defi cient in genetic diseases [46]. Osteogenesis imperfecta is a rare genetic disorder due to abnormal collagen type I production by osteoblasts, resulting in osteopenia, multiple fractures, severe bone deformities and considerably shortened stature. To replace defective osteoblasts, the infusion of allogeneic whole BM or isolated BM-MSCs producing normal collagen type I was evaluated in two studies [47,48]. Although linear growth rate, total body bone mineral content, and fracture rate improved in some patients, the relatively short-term follow-up prevented the authors from drawing fi rm conclusions about the effi cacy of MSC therapy. In a subsequent study with infusions of labelled BM-MSCs, Horwitz and colleagues reported that engraftment was evident in one or more sites, including bone, skin, and marrow stroma, in fi ve out of six patients. Th ese fi ve patients had an acceleration of growth velocity during the fi rst 6 months after infusion [48]. Moreover, the transplantation of allogeneic foetal liver-derived MSCs in a foetus with severe osteogenesis imperfecta led to 0.3% of cell engraftment and diff erentiation of the donor cells into osteocytes until more than 9 months after transplant [49].
Hypophosphatasia, another metabolic bone disease, is a rare, heritable disease due to defi cient activity of tissue nonspecifi c alkaline phosphatase, often causing death in the fi rst year of life due to respiratory complications. In a young girl, transplantation of 5/6 HLA-matched T-celldepleted BM resulted in clinical and radiographic improve ment without correction of the biochemical features of hypophosphatasia during the fi rst 6 months [50]. However, skeletal demineralization occurred 13 months after transplantation and the decision was therefore taken to infuse BM cells that had been expanded ex vivo. Six months later, considerable, lasting clinical and radiographic improvement ensued, still without correction of her biochemical abnormalities. Despite the small number of studies, patients with metabolic bone diseases have benefi ted from allogeneic MSC therapy.

MSCs for the treatment of infl ammatory disorders
Due to their immunosuppressive properties, MSCs may be of interest in the treatment of infl ammatory disorders such as rheumatoid arthritis, which is the most prominent infl ammatory rheumatic disease. To date, confl icting results have been reported using the collagen-induced arthritis (CIA) experimental mouse model. In several studies, the injection of MSCs derived from BM or AT in the CIA mouse model after the establishment of the disease improved the clinical score. Th ese eff ects were associated with a decrease in Th 1-driven infl ammation and TNF-α or IFN-γ serum levels as well as induction of a regulatory T cell phenotype [51,52]. More recently, our group has shown that IL-6-dependent prostaglandin E2 secretion by MSCs inhibits local infl ammation in experimental arthritis [53]. However, this benefi cial eff ect of MSCs in rheumatoid arthritis is still controversial since diff erent studies have shown that the injection of the C3H10T1/2 MSC line, Flk-1(+) MSCs, or MSCs derived from DBA/1 mice did not exert a positive eff ect on CIA or even aggravate the symptoms [54,55]. Th is discrepancy in the eff ect of MSCs may be caused by the diff erent sources of MSCs, but we have reported that altering the course of the disease depends on precise timing of MSC administration [53]. Th is therapeutic window is likely to be associated with the immune status of the mice since it has been recently reported that MSCs are polarized towards an infl ammatory MSC1 or immuno suppressive MSC2 phenotype depending on the type of TLR activation [30].

MSCs for treatment of chronic degenerative disorders
Osteoarthritis is the most frequent rheumatic disease and is characterized by degeneration of articular cartilage, mainly due to changes in the activity of chondrocytes in favor of catabolic activity. However, recent data now suggest that osteoarthritis also involves other joint tissues, with alterations of the meniscus, sclerosis and edema in the underlying subchondral bone as well as intermittent infl ammation of synovium. MSC-based therapy may act via two ways, either preventing cartilage degradation through the secretion of bioactive factors, or by diff erentiating into chondrocytes and contributing to cartilage repair. Th e diff erent options to deliver MSCs to the osteoarthritis joint have been summarized recently [56]. Indeed, the co-culture of human MSCs with primary osteoarthritis chondrocytes allowed the diff erentiation of MSCs towards chondrocytes even in the absence of growth factors. Th is eff ect was dependant on cell-cell communication for secretion of morphogen by chondrocytes, suggesting that MSCs injected in a joint might diff erentiate into chondrocytes [57]. Secretion of bioactive mediators by MSCs may prevent loss of chondrocyte anabolic activity or stimulate progenitors present in the cartilage. As an example, the delivery of autologous MSCs to caprine joints subjected to total meniscectomy and resection of the anterior cruciate ligament resulted in regeneration of meniscal tissue and signifi cant chondroprotection [58]. In an experimental rabbit model of osteoarthritis, transplantation of a hyaluronan-based scaff old seeded with BM-MSCs statistically improved the quality of the regenerated tissue compared to the animal control [59]. Loss of proteoglycans and osteophyte formation were less in the animals treated with MSCs. In humans, eight clinical trials are currently recruiting patients to test the effi cacy of MSC injection for treatment of osteoarthritis. Indeed, a phase I/II trial is currently evaluating the eff ect of MSC injection with hyaluronan (in the form of Chondrogen TM ) to prevent subsequent OA in patients undergoing meniscectomy. Th e mechanisms of MSC-based therapy remain unknown, but it has been speculated that secreted biofactors might reduce fi brocartilage formation or decrease degradation by inhibiting proteinases. Moreover, although osteoarthritis is not considered an infl ammatory disease, secretion of cytokines, namely IL-1β and TNF-α, and immune responses may also be suppressed thanks to the immunomodulatory eff ects of MSCs. Th e various reports therefore argue for a therapeutic effi cacy of MSCs in preventing or limiting osteoarthritis lesions in patients.

Conclusion
Stem cell therapies represent an innovative approach for the treatment of diseases for which currently available treatments are limited. Because MSCs could operate through many diff erent mechanisms, MSC-based therapies are undergoing rapid development and have generated great expectations. Th eir therapeutic potential is currently being explored in a number of phase I/II trials, and three phase III trials have been concluded for the treatment of graft-versus-host-disease, Crohn's disease (Prochymal®, Osiris Th erapeutics) and perianal fi stula (Ontaril®, Cellerix). While the data from numerous clinical trials are encouraging, future studies are obviously needed to confi rm the phase I/II studies. Th ey have nevertheless paved the way for the establishment of feasibility and administration protocols as well as the safety of the procedures. Th is should encourage initiating further clinical studies in non life-threatening pathologies such as rheumatic diseases.

Competing interests
The authors declare that they have no competing interests.