Compared with isolated MSCs, the most important advantage of synovial explants is that the physiological scaffold [extracellular matrix (ECM)] is present around the cells and is provided by the synovial tissue itself. Such scaffold might be more conducive to chondrogenic differentiation than other, non-joint associated scaffolds (or if ECM is absent). Indeed, the synovial tissue is known to differentiate into cartilaginous tissue under both clinicopathological [27] and experimental conditions [15, 28], resulting in the formation of cartilaginous tumors and cartilage-bone-like tissues. Furthermore, our previous studies have shown that bovine synovial explants are able to form more abundant cartilaginous matrix than isolated, alginate-cultured and aggregate-cultured synovial MSCs [29–31]. These findings suggest that a system using synovial explants, which can obviate the need for cell isolation and cell preculturing, could be one of the most promising strategies for the repair of articular-cartilage lesions.
We had shown previously that bovine synovial explants of healthy animals are able to differentiate into cartilaginous tissue after stimulation with appropriate growth factors [15, 16, 18]. In the present study, we demonstrated that synovial explants derived from OA patients have a chondrogenic potential, and maintain this capacity irrespective of age. Given the age-independent high potential of human synovial tissue it may thus be possible to exploit this potential for the repair of cartilaginous lesions in a clinical setting [32].
The activity and the differentiation potential of MSCs of various origins were found to decline with ageing and with the number of passages in vitro [19, 20]. Consequently, the MSCs of elderly donors are considered to be unsuitable for the purposes of tissue engineering [33]. Also the influence of donor age on the chondrogenic potential of MSCs derived from various tissue sources had been addressed by several research groups [34, 35], and the data are conflicting. One possible reason for the discrepant findings might be that MSCs may not be considered as a “universal” population subject to the same process of ageing. The ageing of MSCs may be a tissue-specific phenomenon, or at least a process that is differentially influenced by their origin. Given the great potential value and widespread use of MSCs in regenerative medicine, a clarification of this issue is important, particularly in the fields of orthopaedics and rheumatology, since patients who would qualify for the instigation of an autologous, MSC-based cartilage-repair strategy, would be primarily elderly ones suffering from osteoarthritis.
In a previous study, we demonstrated that synovial cells derived from the knee joints of osteoarthritic patients could be induced to differentiate into cartilage-producing chondrocytes (in vitro) [14], a finding that was reconfirmed by other researchers [36, 37] on similar grounds. However, the influence of donor age on the process of chondrogenic differentiation was not addressed, nor the issue if the generally shifted balance of cell metabolism to catabolic activities of chondrocytes in osteoarthritic joints [38, 39]. These phenomena most likely would affect repair cartilage formation from synovial tissues when originating from such diseased joints; to clarify these potentially adverse issues in tissue engineering of diseased joint cartilage in human patients when using synovial cells and tissue sources of the same joint was the purpose of the present study, using synovial explants.
As evidenced by the deposition of sulfated glycosaminoglycans (volume fraction of metachromasia), BMP-2 alone, and the BMP-2/TGF-β1 combination, induced the chondrogenic differentiation of the synovial explants to the greatest and to similar degrees (no significant difference between the values) (Figures.3A & 3B). Thus TGF-β1 in a combined use with BMP-2 exerted no synergistic/enhancing effect in the human synovial tissue (of OA-joints) differentiation (unlike in synovial tissue of normal bovine joints of young adult animals [18]). Given that the peak volume fraction of metachromasia was approximately 9-fold lower after stimulation with TGF-β1 alone than after exposure to either BMP-2 alone or to the BMP-2/TGF-β1 combination, this is not a surprising finding, and it was, moreover, supported by the immunohistochemical staining profiles for type-II collagen. Although temporal differences in the volume fraction of metachromasia were not significant (owing to high inter-individual variability), a time-dependent increase in this parameter was nonetheless graphically apparent (Fig. 3A), thereby indicating that the chondrogenic activity of the synovial MSCs could be sustained for at least 6 weeks in vitro. When the volume fractions of metachromasia at the 4-week juncture for the two age categories of osteoarthritic patients were compared, no differences were revealed (Fig. 3B). And when this histomorphometric parameter was displayed as a function of an individual’s age, the correlation coefficients were very low (Fig. 4). Thus the general differentiation potential of synovial explants originating from human patients suffering from OA was not impaired as a function of donor age. Moreover the achieved degree of tissue transformation into cartilage-like tissue of the synovial explants was found to be of the same order of magnitude as that encountered in synovial tissue originating from healthy young adult bovine sources [15, 16, 18]. The osteoarthritic process thus seems not to affect the differentiation potential of the synovial tissue in human patients.
The analysis of the gene-expression levels of key anabolic markers of chondrogenesis permits a more discriminative evaluation of the induced differentiation process into cartilaginous tissue.
Type-I collagen is not a marker of chondrogenesis, but a characteristic component of fibrous tissues, and the presence of high levels of its mRNA in repair cartilage is an indication that it is deficient in hyaline-like qualities. Irrespective of the stimulation protocol, the gene-expression levels of type-I collagen were extremely low and bordering on baseline values (Fig. 5). Only after exposure to the BMP-2/TGF-β1 combination a significant difference between the two age categories was revealed, with slightly higher levels in the group of older patients (Fig. 6).
Type-II collagen is a characteristic component of cartilaginous tissues, and, irrespective of the stimulation protocol, its gene-expression levels were higher than for any other marker. However, no age-related differences were revealed (Fig. 6). This is somewhat surprising since on the basis of data in the literature relating to bone marrow derived MSCs and their declining proliferation and differentiation potential with increasing age[19, 33], as well as with perichondrial-derived cells[40] or with periosteum tissue flaps[35], this apparently does not apply for synovial tissue, as found here, nor for isolated synovial-derived MSCs, as previously found [11].
Type-X collagen is a marker of terminal chondrocyte hypertrophy, and high mRNA levels indicate that the extracellular matrix is undergoing calcification, which is undesirable. Irrespective of the stimulation protocol, the gene-expression levels of type-X collagen were elevated, but they were still lower than those of type-II collagen. In the group of younger patients the levels were lower after stimulation with TGF-β1 alone, than after exposure to either BMP-2 alone or the BMP-2/TGF-β1 combination (Fig. 6). Since TGF-β1 was found to suppress the terminal hypertrophic differentiation of chondrocytes [18], this finding is not surprising. Significant age-related differences in the gene-expression levels of type-X collagen were revealed after stimulation with either TGF-β1 alone, or the BMP-2/TGF-β1 combination, the values being higher in the group of older osteoarthritic patients than in the group of younger ones. This phenomenon is in correlation to findings in relation to articular cartilage and osteoarthritic cartilage cells in OA patients [41, 42] and animal models [43], and interestingly is also observed upon differentiation of synovial tissue from osteoarthritic patients; an explanation, however, could not be identified in the literature for this phenomenon. However, given the data relating to the alkaline phosphatase gene activities (see next paragraph) the phenomenon may not be of a generalized metabolic implication but rather a gene-specific activity of type-X collagen gene, without posttranslational activity effects, and thus without functional effects.
Alkaline Phosphatase is likewise a marker of terminal chondrocyte hypertrophy and matrix mineralization [44]. In both age categories of patients, its gene-expression levels were lowest (barely above baseline values) after stimulation with TGF-β1 alone – which accords with the known suppressive effect of this growth factor on terminal hypertrophy –, and lower after exposure to the BMP-2/TGF-β1 combination than after treatment with BMP-2 alone. An age-difference was revealed only after stimulation with the BMP-2/TGF-β1 combination, the values being higher in the group of older osteoarthritic patients than in the group of younger ones.
Type-XI collagen is a marker of cartilage-specific collagen fibril formation since co-expressed with collagen type-II [45]. The gene expression levels of type-XI collagen were highest after exposure to either BMP-2 alone or the BMP-2/TGF-β1 combination. In the latter case, a significant age-related difference was revealed, the values being higher in the group of older osteoarthritic patients than in the group of younger ones. Type-XI collagen gene activities were thus stimulated by both stimulation protocols i.e. by BMP-2 alone or by the BMP-2/TGF-β1 combination, and this was indeed expected given the histological differentiation data obtained and presented above. However the reasons for the differences obtained respecting the different degrees of activation remained unclear.
The gene-expression levels of aggrecan were higher after stimulation with either BMP-2 alone or the BMP-2/TGF-β1 combination than after exposure to TGF-β1 alone. These data accord with the histomorphometric findings for the deposition of sulfated glycosaminoglycans. An age-related difference was revealed only after stimulation with the BMP-2/TGF-β1 combination, the values being higher in the group of older osteoarthritic patients than in the group of younger ones. This data accord with the observations described above relating to type-II collagen gene activities, i.e. the independence of age of the synovial tissue differentiation activity potential.
The gene-expression levels of COMP [46] and Lubricin [47–49], which are markers of the articular-cartilage layer as a whole and of its superficial zone and of synovial fluid, were highest after stimulation with either TGF-β1 alone or the BMP-2/TGF-β1 combination, and lowest after exposure to BMP-2 alone. Indeed, using the latter stimulation protocol, the levels of lubricin were not raised above baseline values. This differs to reported findings relating to muscle-derived MSCs following their chondrogenic differentiation [50]. Significant age-related differences in the gene-expression levels of COMP were revealed after stimulation with either BMP-2 alone or the BMP-2/TGF-β1 combination, the values being higher in the group of older osteoarthritic patients than in the group of younger ones [51]; this possibly may be related to TGF-β1effects [52]. No significant age-related differences in the gene-expression levels of lubricin were disclosed. Irrespective of the stimulation protocol applied, the gene-expression levels of osteocalcin, which is a marker of chondrocytic hypertrophy and mineralization [53], were not raised above baseline values. This finding confirms that the process of chondrogenic differentiation was not proceeding to terminal chondrocyte hypertrophy and matrix mineralization [18] within the time period investigated; however, after longer time periods of culturing these unwanted effects are expected to occur in the absence of TGF-β1[18, 42, 44, 53].
The gene expression levels of matrilin-1, which is a marker of tissue degradation [54], lay below the limits of detection in all instances. Hence, in the newly formed cartilaginous tissue, anabolic processes override catabolic activity [55].
Sox9 is a transcription factor, which acts as a DNA-binding protein during chondrogenic differentiation. Its mRNA levels are believed to peak during the mid-phase of chondrogenesis [56]. Irrespective of the stimulation protocol used, the gene-expression levels of Sox-9 peaked – as expected – at the 4-week juncture. The same temporal pattern of gene-expression was observed for type-X collagen, aggrecan and COMP, which accords with the current opinion that their induction occurs via Sox9 regulated pathways [57].The peak-gene expression levels of Sox9 were higher after stimulation with either BMP-2 alone, or the BMP-2/TGF-β1 combination, than after exposure to TGF-β1 alone, which accords with the histomorphometric findings, namely, that the chondrogenic differentiation of synovial explants occurred to a greater degree after stimulation with either BMP-2 alone or the BMP-2/TGF-β1 combination than after exposure to TGF-β1. Significant age-related differences in the gene-expression levels of Sox9 were revealed after stimulation with either TGF-β1 or the BMP-2/TGF-β1 combination, the values being higher in the group of older osteoarthritic patients than in the group of younger ones.
Compared to the genes for the anabolic markers, those for the catabolic ones [58], which included inflammatory factors [59, 60] [IL-1β, IL-4 (not detected), IL-6, TNF-α (not detected), COX-2 and iNOS] and matrix proteases (MMP-13, ADAMTS-4) were generally expressed at very low levels. As for the anabolic markers, the peak levels were usually attained at the 4-week juncture (Fig. 6). No consistent trend in favor of anyone particular stimulation protocol was observed. But, generally speaking, lower peak levels (the desired results) were achieved in the absence of TGF-β1 (i.e. with BMP-2 alone) than in its presence (TGF-β1, or a combination of BMP-2/TGF-β1). With the exception of IL-1β, no significant age-related differences in the gene-expression levels of the catabolic markers were observed. In the case of IL-1β, the values were slightly higher in the group of older osteoarthritic patients than in the group of younger ones. The finding that the eight catabolic-marker genes were expressed at very low or even non-detectable (IL-4 and TNF-α) levels indicates that the production of a cartilaginous matrix (attested by the histochemical and immunohistochemical observations, and substantiated by the gene-expression levels of the anabolic markers) was positively balanced against degradative processes [59]. The finding that five of the six detected catabolic marker genes were expressed at similar levels in the older and the younger individuals indicates that the state of positive equilibrium between anabolic and catabolic processes was not compromised by ageing.