A potential role for synovial fluid mesenchymal stem cells in ligament regeneration

There is clear evidence for spontaneous joint remodelling and repair in man. Following injury various joint structures, including ligament and tendon, go through a fairly classical cascade of acute injury with immobility, subsequent inflammatory responses and finally adaptive repair response [1]. Synovium also has considerable powers of regeneration given its reappearance following surgical synovectomy [2]. Adaptive responses at fracture sites, at regions of weight bearing and osteophytosis attest to the remodelling capabilities of bone [3]. Additionally, cartilage regeneration at sites of osteophytosis is well recognized and even hyaline articular cartilage-like tissue may occur following corrective osteotomies in OA [4] or following subchondral bone microfracture techniques [5]. A better understanding of the pathophysiological basis for these diverse joint remodelling responses could have major implications for future joint repair strategies. The classical paradigm of joint repair involves the concept of a mesenchymal stem cell (MSC). As MSCs are highly proliferative, clonogenic and multipotential with the ability to give rise to all the joint stromal elements, their key role in tissue-modelling responses seems increasingly likely. The direct study of MSCs in vivo however is still uncommon. The vast majority of functional studies are still performed on culture expanded daughter MSCs, with a resultant colony forming unit-fibroblastic (CFU-F) assay being used to estimate the actual number of native MSC in vivo. The latter analysis has lead to some evidence for the decline in MSC numbers with ageing [6, 7] or also in diseases such as osteoporosis [8]. In recent years, the MSC field has undergone a scientific revolution. Whilst MSCs were originally thought to represent a population of rare bone marrow (BM) stem cells that are mobilized to sites of tissue injury [9], this model is now less popular. To date, virtually every joint structure has been shown to contain resident MSCs including the bone [10], periostium [11], synovium [12], joint fat [13], articular cartilage [14] and even the SF [15, 16]. Considering their very widespread distribution in solid tissues, the presence of MSCs in normal human circulation becomes less relevant [17, 18]. We have previously identified a population of MSCs in the SF of OA and RA subjects [15]. Both SF MSCs and BM MSCs were phenotypically similar and expressed a panel of surface markers (CD105, CD73, D7-FIB, CD271 and others) consistent with their MSC nature. SF MSCs were 20-fold more numerous in OA compared with RA [15]. Furthermore, their numbers appeared to be independent of the degree of joint synovitis raising the possibility that amplification of such cells in OA could be related to attempted joint repair. The original findings on SF MSCs have now been considerably extended. In the study by Morito et al. [19] in the current issue of Rheumatology, the group set out to study the biology of MSCs in vivo following anterior cruciate ligament injury. Although patient numbers were fairly small, the strength of this work from a translational perspective is that the group compare patients with recent cruciate ligament injury and a group of normal controls. The clinical data is supplemented by a microarray analysis of MSCs from different sites and by initial animal model data. The MSC enumeration strategy is based on the traditional CFU-F assay and subsequent extrapolation to the numbers of MSCs in vivo in relationship to post-injury period. Importantly, Morito et al. [19] show that SF MSCs were significantly (100-fold) higher in subjects with ligament injury compared with normal SF. Furthermore, the authors provided initial evidence that SF MSCs were not derived from the BM (as one would have assumed from traditional MSC circulation paradigm), but from the adjacent synovium. Finally, the authors used cell tracking experiments in a rabbit model to furnish evidence that MSCs could contribute to ligament repair via an SF route. The present findings in relationship to ligament injury are truly remarkable when they are interpreted in historical context. For example, a traditional tissue-engineering paradigm assumes the need for a combination of stem cells, scaffolds, growth factors and finally a thorough knowledge of bioengineering principles [20, 21]. However, specifically in relationship to ligaments, the collaboration between Seedhom's group at the University of Leeds and colleagues at Kyoto in Japan showed that detailed attention to only one of these factors could lead to excellent regeneration of ligament in the clinical and experimental settings [22]. Specifically, the implantation of plasma-treated polyester scaffolds in the correct cruciate anatomical orientation and at the correct tension represented a perfectly sufficient criterion for ligament neogenesis with the colonization of the artificial scaffold and the subsequent differentiation of ligament cells [22]. These experiments showed that exogenous growth factors or stem cells were in fact not always needed, but until now it was not clear where reparative stem cells originated from. The traditional paradigm for artificial ligament colonization assumed that MSCs might have migrated from the BM that was exposed during the fixation procedure. The present work by Morito et al., however, shows that following ligament injury, without reported disruption of cartilage integrity or exposure of underlying bone, the numbers of MSCs increase in the SF, suggesting their potential role in ligament repair [19]. Moreover, using a microarray analysis the authors showed that the molecular signature of SF MSCs was more in keeping with synovial origin than with BM origin. The demonstration of intrinsic increase of SF MSCs at sites of ligament injury now retrospectively offers a novel explanation for pioneering artificial ligament scaffold studies and provides a platform for future ligament repair strategies utilizing SF MSCs. The optimal utilization of SF MSCs for ligament repair strategies using gene therapy is also an area that warrants future evaluation [23]. To provide proof-of-principle that SF MSCs could participate in ligament repair, the authors disrupted the cruciate ligament in a rabbit model. Culture-expanded synovial tissue MSCs were labelled with Dil and introduced into the joint. The authors reported preferential adherence of these to sites of damage. It must be pointed out that synovial tissue was used in this setting to facilitate the procurement of adequate numbers of MSCs for the experiment and that these experiments need to be performed with SF MSCs (perhaps in a larger animal model where their adequate numbers could be obtained). Do the present findings merely represent a ligament story or do they have more widespread applicability? As already stated, we have previously found MSCs in arthritic SF and we have recently undertaken further experiments to explore SF MSCs in health and in subjects with the earliest stages of OA [16]. First, we have established the presence of MSCs in normal SF both in man and animals (cows and sheep). Second, we have found that such cells were significantly elevated in subjects with Grades 3–4 chondropathy but where X-rays were deemed to be normal. Third, we found that SF MSCs were functionally and phenotypically different from BM MSCs at the single cell level [16]. These data are in complete agreement with findings by Morito et al. [19]. Moreover, we found that single SF MSCs were consistently chondrogenic in vitro compared with BM MSCs, which showed very variable chondrogenesis. The origin of SF MSCs in subjects with damaged ligaments or injured cartilage is still not completely established. In agreement with Morito et al. [19], our data suggest a possible synovial origin. We noted the presence of synovial tissue clumps in the fluid, which closely correlated with SF MSC numbers [16]. Indeed, the presence of synovial clumps in SF has been observed previously [24]. In their elegant early work with rabbits, Hunziker and Rosenberg [25] provided evidence for regenerative cells directly migrating from synovium into partial thickness chondral defects. However, the rabbit model cannot be directly extrapolated to man since the latter has a synovial membrane directly adhered to the weight-bearing cartilage surface, which is not the case for humans. To further extend these ideas, our data showed that SF could represent more than a mere conduit for synovial MSC passage, since SF had a direct growth-promoting effect on these cells [16]. Therefore, based on presently available data it seems feasible to coax SF MSCs to colonize either artificial ligament scaffolds and possibly damaged or incompletely ruptured ligaments to affect tissue repair responses. However, given that there is a lack of evidence for spontaneous repair of partial thickness chondral defects it is less clear that a similar scenario may be operational with respect to articular cartilage in mature adult joints. Nevertheless, this remains a possibility since even in advanced OA some evidence for cartilage repair has been furnished [4]. There is a current controversy on the precise relationship between MSCs and ‘transformed’ and invasive synovial fibroblasts in RA. This is because different groups have used fairly similar isolation techniques (based on plastic adherence and culture expansion) but studied different characteristics of these cells. Gay and colleagues [26] previously showed that ‘floating’ fibroblast-like cells from SF placed in a SCID mouse model were capable of destroying cartilage. Therefore, it is worth bearing in mind that in highly inflamed joints SF MSCs may also contribute to joint damage and further work is needed to explore these possibilities. As already explained, it is possible to repair human ligaments without the use of exogenous growth factors or stem cells. The present study by Morito et al. [19] underscores the need to understand adaptive stem cell responses in vivo and further indicate that such improved biological understanding could ultimately lead to new treatments for joint disease. The augmentation of SF MSC adhesion, proliferation and differentiation could have major implications for ligament repair strategies. Rheumatologists who have a rapidly increasing expertise with the use of small molecules and biological agents need to be at the vanguard of such developments since the therapeutic manipulation of MSCs in vivo could play a major role in the potential success of such strategies. To summarize, complete transaction of a ligament cannot repair spontaneously due to loss of the scaffold necessary for ligamentogenesis. The present findings suggest that partial injury to cruciate ligaments may be associated with repair responses, at least in part, mediated by SF MSCs and that the repair response may not be confined to cells intrinsic to the ligament itself. Further experimental work is needed to establish the origin and role of SF MSCs in joint repair in larger number of cases. Given the emerging importance of ligaments in the phenotypic expression of OA at diverse sites including the hands and knees, it is possible that maladaptive ligament tissue repair responses are pivotal for disease expression in a wide array of settings and not just for traumatic ligament injuries [27]. To conclude, the present work by Morito et al. [19] represents an exciting development and could open up a new era into understanding the role of stem cells in ligament biology in health and disease. Disclosure statement: The authors have declared no conflicts of interest.