Transforming growth factor beta 3 involved in the pathogenesis of synovial chondromatosis of temporomandibular joint

Synovial chondromatosis (SC) of temporomandibular joint is rare proliferative disorder featured by the formation of cartilaginous nodules in synovium and joint space. Transforming growth factor beta 3 (TGF-β3) is closely related to chondrogenic differentiation, and might participate in pathogenesis of SC. We discovered that increased quantity of synoviocytes and blood vessels were observed in SC synovium. The vessel wall and sublining fibroblasts were stained positively by the antibodies against TGF-β3, fibroblast growth factor 2 (FGF-2), and CD34. In loose bodies (LBs), TGF-β3 was mainly expressed in chondrocytes and FGF-2 was expressed in chondrocytes, fibroblasts, and vessel walls. Expressions of TGF-β1, TGF-β3, FGF-2, Sox9, Wnt-4, Foxc2, and VEGF-A mRNA were significantly higher in SC synovium. Stimulation of TGF-β3 on synoviocytes increased alkaline phosphatase (ALP) activity and expressions of chondrogenic genes (Sox9, Col2α1, Aggrecan, Wnt-4, and Wnt-11), osteogenic genes (Runx2, Foxc2, osteocalcin, and Col1α1), and VEGF-A, but failed to influence FGF-2 expression. However, the addition of FGF-2 increased TGF-β3 expression. In conclusion, TGF-β3 existed in synovium and LBs of SC, and was responsible for the pathogenesis of SC.

Synovial chondromatosis (SC) of temporomandibular joint is rare proliferative disorder featured by the formation of cartilaginous nodules in synovium and joint space. Transforming growth factor beta 3 (TGF-b3) is closely related to chondrogenic differentiation, and might participate in pathogenesis of SC. We discovered that increased quantity of synoviocytes and blood vessels were observed in SC synovium. The vessel wall and sublining fibroblasts were stained positively by the antibodies against TGF-b3, fibroblast growth factor 2 (FGF-2), and CD34. In loose bodies (LBs), TGF-b3 was mainly expressed in chondrocytes and FGF-2 was expressed in chondrocytes, fibroblasts, and vessel walls. Expressions of TGF-b1, TGF-b3, FGF-2, Sox9, Wnt-4, Foxc2, and VEGF-A mRNA were significantly higher in SC synovium. Stimulation of TGF-b3 on synoviocytes increased alkaline phosphatase (ALP) activity and expressions of chondrogenic genes (Sox9, Col2a1, Aggrecan, Wnt-4, and Wnt-11), osteogenic genes (Runx2, Foxc2, osteocalcin, and Col1a1), and VEGF-A, but failed to influence FGF-2 expression. However, the addition of FGF-2 increased TGF-b3 expression. In conclusion, TGF-b3 existed in synovium and LBs of SC, and was responsible for the pathogenesis of SC. S ynovial chondromatosis (SC) of temporomandibular joint (TMJ) is a rare proliferative disorder of synovium accompanied by formation of cartilaginous nodules in synovium and joint space 1-3 , as well as secondary calcification and ossification 4 . The clinical manifestations include unilateral pain, swelling, clicking, occlusal changes, crepitation, and limited mandibular function 5,6 .
Transforming growth factor b3 (TGF-b3) was reported to be a potent mediator for inducing chondrogenesis of mesenchymal stem cells (MSCs) [7][8][9] , and to increase the production of cartilaginous extracellular matrix (ECM) 7,10 . The synoviocyte in synovium of TMJ is recognized as MSC because of its potentials to differentiate into adipocyte, chondrocyte, and osteocyte lineages 11,12 . Therefore, we hypothesized that TGF-b3 might be responsible for chondrogenic differentiation of TMJ synoviocytes and the pathogenesis of SC.
Besides, our group previously discovered that fibroblast growth factor 2 (FGF-2) was responsible for the formation of loose bodies (LBs) and increased blood vessels of synovium. The relation between TGF-b3 and FGF-2 was therefore investigated in this research.
The main purpose of this study was to investigate the roles of TGF-b3 in the formation of LBs and its relation with FGF-2.
During the surgery, cartilaginous nodules which are also called LBs were observed in the synovium and articular cavity of TMJ. Most of the LBs were too stiff to be chopped by scissors. Two types of LBs were found based upon histological features. LB of the first type was composed of single cartilaginous nodule ( Fig. 2a-d). The second type LB was formed by a number of small cartilaginous nodules ( Fig. 2eh). In the first type, a thick synovium which contained increased amount of synoviocytes and blood vessels was found covering the LB ( Fig. 2a and b). In the second type, a connective tissue containing small blood vessels and fibroblasts was found to separate the numerous cartilaginous nodules ( Fig. 2e and f). TGF-b3 was expressed mainly in the chondrocytes ( Fig. 2c and g). FGF-2 was found in chondrocytes of LB ( Fig. 2d and h), vessel wall of the synovium (Fig. 2d) and fibroblasts of the connective tissue (Fig. 2h).
Reverse transcription PCR (RT-PCR) for analyzing SC and normal synovium. The expressions of TGF-b1 and TGF-b3 in SC synovium were 1.8 and 60.7 times that of the normal synovium ( Fig. 3), respectively. In regard to chondrogenic genes, the mRNA expressions of Sox9 and Wnt-4 in SC synovium were 4.3 and 9.7 times that of the control (Fig. 3), respectively. In regard to osteogenic genes, Foxc2 expression in SC synovium was 4.7 times that of the control (Fig. 3). However, the Runx2 expressions between SC and control synovium had no difference significantly (Fig. 3). The expressions of FGF-2 and vascular endothelial growth factor A (VEGF-A) in the SC synovium were 4.1 and 15.9 times that of the control (Fig. 3), respectively.
The addition of FGF-2 in the culture medium for SC synoviocytes increased the TGF-b3 gene expression by 16.9-fold (Fig. 5). However, the addition of TGF-b3 did not influence the FGF-2 expression significantly (Fig. 5). The alkaline phosphatase (ALP) activities in the two groups with and without the addition of TGF-b3 were 18.1 and 9.6 U/gprot (Fig. 5), respectively.

Discussion
As the TMJ SC is a rare disease, most studies in literature are case report. Basic investigations of SC focus mainly on certain cytokines or proteins in synovium and LBs 1,6,13-15 through analysis of IHC staining methods. Comprehensive studies on this disease are hard to find.
This study focused on the role of TGF-b3 in the pathogenesis of SC, because TGF-b3 was potent mediator for inducing chondrogenesis of MSCs 7,8,16 and had a higher chondrogenic potential of a more rapid differentiation than TGF-b1 17 . Besides, the cartilaginous nodules were observed in both synovium and joint space 4,5,18 , and made of cartilaginous ECM and chondrocytes which are absent in normal synovium 19,20 . Therefore, we speculated that TGF-b3 might exist in the synovium and LBs. In this study, the existence of TGF-b3 was confirmed by the IHC assay (Fig. 1c, Fig. 2c and g) and the RT-PCR for synovium of SC (Fig. 3).
Then this study examined the effects that TGF-b3 exerted on SC synoviocytes. First, TGF-b3 induced chondrogenic differentiation of SC synoviocytes, supported by the increased levels of mRNA expressions of Sox9, Col2a1, aggrecan, Wnt-4 and Wnt-11 in this study (Fig. 4). Sox9 is a potent mediator of chondrocyte phenotype, and regulates the expressions of key  Besides, TGF-b3 elevated the expressions of osteogenic genes, such as Runx2, Foxc2, Col1a1 and osteocalcin in this study (Fig. 4).
Foxc2 is involved in regulating both osteogenesis and angiogenesis of MSCs 23 . The angiogenic effect of Foxc2 might explain partly for the phenomenon of increased angiogenesis in synovium of SC (Fig. 1b,  Fig. 2b). Runx2 is a common target of bone morphogenetic protein-2  (BMP-2) and TGF-b1, and plays an essential role in osteogenic differentiation 24 . BMP-2 was reported to participate in the pathogenesis of cartilaginous and osteogenic metaplasia in SC 25 . However, the Runx2 expression had no statistical difference between SC and control synovium (Fig. 3), a phenomenon which could be explained by our group's previous finding 26 that the stimulation of FGF-2 on SC synoviocytes down-regulated the Runx2 expression. Therefore, the integrated effect of FGF-2 and TGF-b3 caused no statistical difference in Runx2 expression between SC and control synovium.
Moreover, TGF-b3 upregulated the expression of angiogenic gene VEGF-A. Previously, we found that the addition of FGF-2 to the medium for SC synoviocytes also upregulated the expression of VEGF-A 26 . Therefore, both TGF-b3 and FGF-2 participated in the angiogenesis of SC pathogenesis. The higher expression of VEGF-A accorded with the phenomenon of increased quantity of blood vessels in SC synovium (Fig. 1b, Fig. 2a and b), as VEGF-A is wellknown to be mitogenic for endothelial cells 27 and to regulate vascular permeability 28 .
Then, we investigated the interaction between TGF-b3 and FGF-2. Interestingly, TGF-b3 could not influence the FGF-2 expression but FGF-2 could upregulate the TGF-b3 expression. Previously, we showed that the stimulation of FGF-2 on SC synoviocytes upregulated the expressions of chondrogenic and osteogenic genes 26 . Therefore, we concluded that both TGF-b3 and FGF-2 were involved in the pathogenesis of SC, and that the production of TGF-b3 could be regulated by FGF-2.
The clinical importance and guiding significance of this study should be emphasized. For clinicians, this study provided a possible way to treat SC in early stage. By intra-articular injection of the medicine that can suppress the production and activity of TGF-b3, the cartilaginous ECM accumulation process may be stopped. By intra-articular injection of the medicine that can suppress angiogenesis, the nutrient supply may be cut off. Therefore, using these two ways, the formation of LBs may be restrained. For researchers, this study demonstrated the basic structures and features of the synovium and LB of SC. Future researchers may follow the ideas of chondrogenic differentiation of MSCs and the angiogenesis to choose related new cytokines or proteins to make further investigations. Researchers of cartilage tissue-engineering field may also be benefit from the investigations of SC, because by using the chondrogenic cytokines involved in the pathogenesis of SC, it may be possible for them to engineer the artificial cartilage with more similar biological and mechanical properties with natural cartilage.  of classic treatments for histological observations, the paraffin-embedded sections of 4 mm-thick were acquired, and disposed with haematoxylin and eosin (HE) staining.
RT-PCR for comparing SC and normal synovium. To investigate the mRNA expressions of TGF-b1, TGF-b3, FGF-2, VEGF-A, Sox9, Wnt-4, Runx2 and Foxc2, RT-PCR assay was performed. The primer sequences of the above targets were listed in Table 1. As described previously 30,31 , SC and control synovium specimens were powdered in liquid nitrogen. According to the instructions of the manufacturer, trizol reagent (Invitrogen, CA, USA) was used to extract the total RNA from the specimens. Reverse transcriptase and oligo dTs were used to achieve synthesizing the cDNA with the assistance of ReverTra Ace kit (Toyobo, Osaka, Japan). The RT-PCR assay was conducted in triplicate using the following protocols: pre-incubation for 1 min at 95uC, followed by 40 PCR cycles for 15 s at 95uC, 20 s at 58uC, and 20 s at 72uC. Electrophoresis was performed subsequently in 1.5% agarose gel containing ethidium bromide. The products were visualized under ultraviolet light, and photographed. The densitometric value of DNA bands was measured with the image analysis software (NIH Image J).
Cell culture. Cell culture was performed as described previously 32 . Briefly, the SC and control synovium specimens were washed with phosphate buffered solution (PBS) containing penicillin and streptomycin. Then the synovium specimens were cut into 1 mm 3 , and maintained with a solution containing dulbecco's modified eagle medium (DMEM, SH30022.01B, HyCloneH) and 15% fetal bovine serum (FBS, SV30087,HyCloneH) in a humidified atmosphere of 5% CO 2 and 95% air at 37uC. When reaching confluence, the synoviocytes were dissociated by trypsin (SH30042,HyCloneH). Then the synoviocytes were cultured with DMEM containing 10% FBS.
Realtime PCR for evaluate the effects of TGF-b3 on SC synoviocytes. The SC synoviocytes of passage 3 rd to 6 th were cultured with DMEM containing 10% FBS in a 6-well plate. After reaching 70% confluence, the SC synoviocytes were maintained in the DMEM containing 4% FBS in presence or absence of 10 ng/ml recombinant human TGF-b3 (100-36E, PeproTech, USA) for 6 days. The medium was replaced every two days. Realtime PCR was performed to investigate the gene expression levels of Col2a1, Aggrecan, Sox9, VEGF-A, Wnt-4, Wnt-11, Runx2, Foxc2, Col1a1, and FGF-2. The primer sequences of the targets were listed in Table 1.
As described previously 31 , according to the instructions of the manufacturer, trizol reagent was used to extract the total RNA from the specimens. The RevertAid First Strand cDNA Synthesis Kit (K-1622, ReverTra Ace-a) was used to achieve reverse transcribing total RNA into cDNA. The semi-quantitative PCR of cDNA samples were disposed with TOYOBO THUNDERBIRD SYBR qPCR Mix (QPS-201). The realtime PCR assay was conducted in triplicate using the following protocols: preincubation for 1 min at 95uC, followed by 40 PCR cycles for 15 s at 95uC, 20 s at 58uC, and 20 s at 72uC. The method of comparative threshold cycle (DDCt) was used to evaluate the expression levels of the desired genes which was normalized by the expression level of b-actin gene measured in each sample.
The stimulation of FGF-2 on TGF-b3 expression. After reaching 70% confluence, the SC synoviocytes were maintained in the DMEM containing 4% FBS in presence or absence of 10 ng/ml recombinant human FGF-2 (AF-100-18B, PeproTech, USA) for 6 days. Then RNA was extracted and realtime PCR was applied to investigate the expression level of TGF-b3 mRNA.
ALP activity measurement. SC synoviocytes were cultured in DMEM containing 4% FBS with or without 10 ng/ml TGF-b3 for 6 days. As previously described 33 , the SC synoviocytes were lysed in 1% Triton X-100. The ALP activity was measured with colorimetric ALP Assay Kit (A059-2, Nanjing Jiancheng Bioengineering Institute) in accordance with the instructions of the manufacturer, and standardized to the concentration of total protein by bicinchonininc acid (BCA) Assay (P0012, Beyotime Institute of Biotechnology).
Statistical analysis. Statistical analyses were conducted using SPSS 11.0.0. All data in this study were shown as mean 6 standard deviation. Paired t-test was applied to evaluate the significance. a 5 0.05 was designated as the significance levels.

Conclusion
TGF-b3 was expressed in the synovium and LBs of TMJ SC. TGF-b3 upregulated the expressions of chondrogenic, osteogenic, and angiogenic genes in SC synoviocytes. The TGF-b3 expression was upregulated by FGF-2. Therefore, TGF-b3 was responsible for the pathogenesis of TMJ SC. Inhibition of the functions of TGF-b3 might be a possible method to control the formation and growing of the cartilaginous nodules in early stage of SC in clinical treatment. By verifying the conjecture that the formation of cartilaginous nodule in synovium of TMJ was the outcome of the chondrogenic differentiation of the synoviocytes, this study provided a basic explanation for this rare clinical disease, and also provided an understanding of cartilage regeneration as well as the blood supply for tissueengineering field. In future studies, we will do investigations to facilitate the formation of cartilaginous nodules in animal TMJ in order to reflect and further discover the process of pathogenesis of SC.