Differentiated properties of SMSCs compared with other MSCs SMSCs have a greater proliferative capability and colony-forming potential than other stem cell sources in vitro, including bone marrow, periosteum, muscle and fat. In the early years, analysis by De Bari and colleagues [2] had already shown the superior proliferative abilities of SMSCs, which maintained growth rates over 30 population doublings, and demonstrated multilineage differentiation capacities. Recently, Yoshimura et al [3] compared the proliferation of rat mesenchymal stem cells derived from the bone marrow, synovium, periosteum, adipose tissue and muscle: implanted with the same amount of cells at passage 1, the SMSCs showed the greatest proliferative rate among them, with the largest amount of cloning and the cells per clone. Jo et al [4] demonstrated the colony-forming potential per nucleated cell in SMSCs to be 1 in 12.5-80 in comparison with 1 in 103-104 in BMSCs. Sakaguchi et al [5] reported that plating at optimal density for 14 days came out with an average of 21 000 cells per 1 mg of synovium. They also found that SMSCs retained the proliferative capacity over four passages, whereas muscle MSCs and adipose MSCs lost their proliferative capacity over four and seven passages. Nimura et al [6] compared the proliferation of human SMSCs and human BMSCs with autologous human serum. They found that human SMSCs expressed higher levels of platelet-derived growth factor (PDGF) receptor α, than did human BMSCs. Neutralizing PDGF, which was contained highly in human serum, decreased the proliferation of human SMSCs.   Although SMSCs are distinct from other MSCs, their surface epitope expressions are similar. Particularly, there is very little difference between surface epitopes for SMSCs and those for BMSCs. Several studies have shown that both cells types are positive for CD44, CD73, CD90 and CD105Hermida-Gómez et al [11] analyzed SMSCs in the membrane of osteoarthritis patients using flow cytometry, and discovered that cells expressed both CD44 and CD105 more than CD44 and CD90, CD 90 and CD105. The cells expressing SMSCs surface marker mainly distribute in the lining layer of synovial membrane, with very few in other region. Interestingly, Arufe et al [13] studied the subpopulations of SMSCs enriched for CD73, CD106, and CD271 markers, which had different profiles of cells positive for the MSC markers CD44, CD69, CD73, CD90, and CD105., but negative for CD14, CD34, and CD45. Among them, only CD90 is found to be significantly higher in mean fluorescent intensity in SMSCs than BMSCs, which may explain the increased chondrogenic potential of SMSCs[7-8]. SMSCs are also positive for CD9, CD10, CD13, CD54, CD55, CD166, and D7-FIB, but they are not unique to SMSCs[7]. Subsequently, studies have found that SMSCs can abundantly express CD49d, but BMSCs do not[9]. It has been reported that SMSCs can retain the surface epitope markers of CD73, CD90 and CD105, after six passages[10]. Another study reveals that it can retain not only the expression of CD90 (endotheliocyte marker)[11] but also the expression of D7-FIB (fibroblasts marker)[12] after the subculture.   Chondrogenesis of SMSCs compared with other MSCs Multiple studies have demonstrated the greatly chondrogenic capacity of SMSCs among MSCs. As cell-to-cell contact may promote chondrogenesis and inhibit osteogenesis in cells[14], pellet culture has been used in many studies. Sakaguchi et al [5] reported that using pellet culture multiple rat MSCs, the pellets larger than 1 mm were shown in SMSCs, BMSCs and periosteum MSCs cultures. In contrast, pellets in muscle and adipose MSCs were both smaller than 1 mm. The pellets from SMSCs and periosteum MSCs were also heavier than those from BMSCs, muscle and adipose MSCs. What’s more, in SMSCs, synthesis of chondroitin sulfate and hyaluronan was highest.   There are more in-vitro studies than in-vivo studies to examine the chondrogenesis of SMSCs. Koga et al [15] found that SMSCs, seeded in collagen gel, showed an enhanced chondrogenic capacity in rabbit osteochondral defects compared with other MSCs. Significant cartilage matrix and well histological score were found in defects with SMSCs and BMSCs. In contrast, defect repaired with adipose and muscle MSCs were badly healed[5]. There are several problems worth considering in the application of SMSCs, importantly, such as dedifferentiation, degeneration and ossification of hyaline cartilage. Recent studies have showed that the capacity of undergoing osteogenesis in SMSCs seems to be reduced compared with that in other MSCs. SMSCs have been reported by Koga et al[16] to retain metachromatic staining at least for 24 weeks in repaired cartilage and do not show signs of ossification. Shirasawa et al [14] found that, as the culture time of SMSCs increased, endogenous BMPs and BMP receptors were not detected or decreased. Throughout the culture, expression of the osteogenic marker runx-2 did not change. The pellet did not reveal signs of hypertrophy or bone formation, though some osteogenic protein was detected. Another important appearance is that only a small amount of bone sialoprotein mRNA was expressed by the pellets in SMSCs, and neither did osteocalcin mRNA nor signs of calcification.   Suitable culture conditions for SMSCs in cartilage repair Growth and differentiation factors The most common growth and differentiation factors used in culturing SMSCs recently include transforming growth factor-β (TGF-β) superfamily, basic fibroblast growth factor (bFGF), bone morphogenetic protein 2 (BMP-2). These factors, especially, the synergistic action of TGF-β and BMP-2, have been proved to be able to stimulate the chondrogenesis of SMSCs. Fan and colleagues engineered cartilage using SMSCs in vitro with injectable gellan gum hydrogels. The cells were divided into three groups, culturing in TGF-β1, TGF-β3 and BMP-2, respectively. After 42 days, the two TGF-β groups showed more significant chondrogenesis compared with the BMP-2 group[17]. Subsequently, Rui et al[18] used BMP-2 to promote the TGF-β3-induced chondrogenesis of SMSCs isolated from human osteoarthritic synovium in a pellet culture system. Compared with the TGF-β3-alone group, cell pellets treated with TGF-β3 and BMP-2 showed a significantly better result in diameter, weight, glycosaminoglycans (GAGs), the level of collagen type II and other chondrogenic markers, except COL10A1[18]. Recently, Shintani et al [19] introduced 10 ng/mL TGF-ß1 into the BMP-2-induced bovine synovial explants that were furnished with SMSCs for the cartilage repair, compared with the same density of FGF-2. The FGF-2 group only increased metachromatic staining for GAGs during the first week of culturing alone, while the TGF-β1 group not only more greatly increased the metachromasia, but also enhanced the biochemically-assayed accumulation of GAGs, when it was presented throughout the entire culturing period. Moreover, at the same time, TGF-β1 arrested the downstream differentiation of cells at an early stage of hypertrophy. Kim et al [20] also applied various concentrations of bFGF to monolayer cultures of SMSCs. As a result, SMSCs showed augmented sizes, weights, and GAG accumulation of pellets by bFGF supplementation. Up-regulated messenger RNA and protein expression of type II and type X collagens were also detected. The 10 ng/mL of bFGF was elected as the optimal concentration. Meanwhile, cell shrinkage and increased actin expression were noted.   It is worthwhile to take notice that the TGF-β pathway to induce the chondrogenesis of SMSCs has been reported already. In 2012, Xu et al [21] described RhoA/Rho kinase signaling as the regulation of TGF-β1-induced chondrogenesis and actin organization of SMSCs, interacting with the Smad pathway. Primary isolated SMSCs were treated with TGF-β1. C3 transferase, Y27632 and SB431542 were used to evaluate the function of RhoA/Rho kinase and Smads as the specific biochemical inhibitors. As a result, the activation of the RhoA/Rho kinase pathway was in response to the stimulation of TGF-β1 in SMSCs, and concomitantly cytoskeletal reorganization was induced, which was specifically blocked by C3 transferase and Y27632. In addition, the phosphorylation of Smad2/3 was reduced by Y27632 treatment, in a concentration-dependent manner. This may be an important discovery for the successful utilization of SMSCs to achieve articular cartilage tissue engineering.   Other influential factors Numerous studies about the valuable factors that could regulate the chondrogenesis and proliferation of SMSCs have been reported in the past few years, many of which showed new train of thought to the culturing. Shimaya            et al [22] added magnesium into the SMSCs-based culturing integrin, and performed assays for adherence and chondrogenesis of SMSCs in vitro and in vivo in a rabbit osteochondral defect model: through the study, the adhesion of human SMSCs to collagen was increased by magnesium, meanwhile inhibited by neutralizing antibodies for integrin α3 and β1. And it promoted synthesis of cartilage matrix during in vitro chondrogenesis of SMSCs, which was diminished by neutralizing antibodies for integrin β1 but not for integrin α3. Positive effects were also found in adherence of human SMSCs to osteochondral defects ex vivo, adherence and cartilage formation of SMSCs in vivo in rabbits[22]. Another research reported by Shimomura and co-workers[23] in 2010 focused on the influence of skeletal maturity. They compared the repair quality of damaged articular cartilage between immature and mature pigs, which was explanted by a scaffold-free three-dimensional tissue-engineered construct derived from SMSCs. No skeletal maturity-dependent difference was found in proliferation or chondrogenic differentiation capacity of the porcine SMSCs, suggesting the feasibility of allogenic SMSC-based cartilage repair over generations. Shintani  et al [24] investigated the effects of dexamethasone upon the TGF-β1- and BMP-2-induced chondrogenesis of different MSCs and microenvironments: in aggregates of SMSCs, DEX exerted remarkable effect on neither TGF-β1 nor BMP-2-induced chondrogenesis. In synovial explants, it inhibited BMP-2-induced chondrogenesis   greatly, but had little impact on the TGF-β1-induced response, showing the fact that steroids are not always necessary for the chondrogenesis of MSCs in vitro. In 2012, Bertram et al [25] spanned a range of osmolarities (264-375 mOsm) in the SMSCs-culture media to see the chondrogenic changes of SMSCs isolated from the synovial fluid of normal people or osteoarthritis and rheumatoid arthritis patients. Consequently, SMSCs from arthritic joints showed decreased chondrogenic potential compared to SMSCs isolated from normal synovial fluid. Under the osmolarity conditions for which they were initially derived within, the SMSCs retained increased chondrogenic potential.   Interestingly, the way to culture SMSCs can also influence the property of SMSCs. In 2012, Suzuki et al [26] compared the aggregate-cultured SMSCs and the monolayer- cultured SMSCs both isolated from human and rabbits. In in-vitro studies, the human SMSCs derived from aggregates showed higher amounts of cartilage matrix synthesis in pellets than that cultured in a monolayer. In in-vivo studies in rabbits, aggregates of SMSCs could adhere rapidly on the osteochondral defects by surface tension, and furthermore, stay without any loss.   Culture serum The most common serum used in the SMSCs culturing in recent years contains the autologous human serum, fetal bovine serum (FBS). Nimura et al [6] cultured synovial and bone marrow MSCs obtained from 18 donors with autologous human serum or FBS. Notably, human SMSCs expanded more in human serum than in FBS through PDGF, whereas the opposite results were obtained with BMSCs. It also showed that neutralizing PDGF, which receptor is higher in SMSCs than BMSCs, decreased the proliferation of SMSCs with autologous human serum. Later, their group expanded human SMSCs with either autologous human serum or FBS after heat-inactivation. It is demonstrated that heat-inactivated autologous human serum enhanced proliferation of SMSCs. Moreover, heat-inactivation of each types of serum did not affect calcification of SMSCs. And the adipogenesis of SMSCs with autologous human serum was similar to or less than that with FBS[27].   Scaffolds The culture scaffold selection is also becoming the study heat point in these years. In 2010, Fan et al [17] encapsulated passage-4 monolayer-amplified rabbit SMSCs in injectable gellan gum hydrogels as the culture scaffold followed by culture within TGF-β1, TGF-β3 or BMP-2 for 42 days. The viability of SMSC in hydrogels in these three groups remained high at culture time, demonstrating the potential application in clinical cartilage repair. The researchers also applied the non-degradable poly(ethylene glycol) diacrylate-based hydrogel and biodegradable phosphoester-poly(ethylene glycol)-based hydrogel as three-dimensional scaffolds encapsulated with rabbit SMSCs, and compared the viability of SMSCs in both hydrogels. It is reported that SMSCs continue to have a high viability, and positive SMSC chondrogenesis is successfully achieved in both gels by induction of TGF-β1 or TGF-β3, with the best outcome in the poly(ethylene glycol) diacrylate system[28]. In 2011, Qi et  al [29] seeded CD105-positive rat SMSCs onto the chitosan-alginate composite three-dimensional porous scaffolds, which was put into chondrogenic culture medium within TGF-β3 and BMP-2 in vitro. After 2 weeks, cells attached and proliferated well on scaffolds, and secreted extracellular matrix. The sorted cells cultured in scaffolds demonstrated more chondrogenic differentiation as well, compared with non-sorted cells. In 2013, Zhang et al[30] applied type I collagen scaffold with stromal cell-derived factor-1 to create a matrix environment for SMSCs-based cartilage repair. Through ex vivo and in vitro studies, subchondral bone or type I collagen scaffold was more permissive for SMSCs adhesion than cartilage or type II collagen scaffold.   SMSCs in treatment of other cartilage diseases Through the literatures available so far, it is certain that diseases around the cartilage can influence the characteristics of SMSCs. Sekiya et al[31] isolated SMSCs from knee joints of normal people, anterior cruciate ligament injury, mild osteoarthritis and severe osteoarthritis patients, to analyze the relationship between MSCs number in the synovial fluid and the different kinds of knee joint. Through arthroscopy, the MSCs in the synovial fluid were hardly observed in normal volunteers, but the number increased as the grading of osteoarthritis goes up. Jones et al [32] evaluated the frequency and multipotency of SMSCs from rheumatoid arthritis patients in relation to synovial inflammation measured using the arthroscopic visual analogue score. It came out with the fact that the chondrogenesis of SMSCs was inhibited in direct relation to visual analogue score and reduced compared with control SMSCs from osteoarthritis patients. They concluded a negative relationship between SMSCs chondrogenic and clonogenic capacities and the joint inflammation in rheumatoid arthritis. Morito-100 ng/mL interleukin-1β, the resultant size of pellets from normal ovine SMSCs has been significantly inhibited. et al [33] obtained human synovial fluid from anterior cruciate ligament injury patients around the time of reconstruction surgery and from healthy volunteers, then cultured and examined the MSCs in the synovial fluid. Through the study, the number of MSCs in the synovial fluid from intra-articular ligament injury patients was demonstrated to be 100 times more than those from healthy volunteers. No significant differences in surface epitope and differentiation potential were observed among the cells around the time of the surgery. Ando et al [34] isolated SMSCs from six sheep at 2 weeks after experimental anterior cruciate ligament core surgeries: The size of SMSC pellets from operated knees after culturing was significantly smaller than that from contralateral knees. With the addition of 1   Studies in vivo for cartilage repair Articular cartilage repair with SMSCs has been reported with an increasing number of animal studies in vivo in these few years, though clinical studies are still largely lacking. Currently it can help to repair cartilage defects to a limited extent in vivo. Koga et al [16]created full-thickness osteochondral defects in the knees of adult rabbits, which were filled with labeled SMSCs and covered with periosteum. It was demonstrated that SMSCs were able to form a cartilage matrix with an abundance of sulfated GAGs and type II collagen, and regionally specified in accordance to local microenvironment. Up to 24 weeks, integration with native tissue was noted, while the thickness of the repair tissue decreased. Nimura et al[6] reported a study using SMSCs transplanted into full-thickness cartilage defects of the knees in rabbits, with a comparison between the autologous serum cultured group and the FBS group. The effect revealed that chondrogenic potential of rabbit SMSCs in vivo was similar in both two groups. Pei et al [35] implanted rabbit SMSCs after 1-month incubation into the knees to repair the full-thickness osteochondral defects, along compared with the scaffold group (fibrin glue-saturated PGA) and the empty group (untreated). Six months after implantation, cartilage defects were full of smooth hyaline-like cartilage with no detectable macrophages and type I collagen but a high expression of GAG and type II collagen, which were integrated with the surrounding native cartilage as well, compared with the scaffold and empty groups respectively resurfaced with fibrous-like and fibrocartilage tissue. Zhang et al [30] implanted rabbit SMSCs encapsulated in type I collagen scaffold with cell-derived factor-1 into the rabbit to repair partial-thickness defects. Good results were shown that in situ self-repair of partial-thickness defects enhanced 6 weeks post-injury. It also brought out the conclusion that in partial-thickness defects, the inferior self-repair capacity is partially owing to the non-permissive microenvironment. Ando et al [36] implanted the in vitro generated scaffold-free tissue-engineered construct derived from porcine SMSCs into the cartilage defects in the medial femoral condyle of 4-month-old pigs. After 6 months, the tissue-engineered construct repaired tissue showed mechanical properties similar to the normal porcine cartilage, as well as secure biological integration to the native cartilage. Nakamura    et al[37] injected porcine SMSCs into the knees with artificial cartilage defects in pigs, compared with the non-treated control knees. Quantification analyses for arthroscopy, histology and MRI all revealed a better outcome in the SMSC-treated knees than in the non-treated control knees."