Promoting Effect of Basic Fibroblast Growth Factor in Synovial Mesenchymal Stem Cell-Based Cartilage Regeneration

Synovial mesenchymal stem cell (SMSC) is the promising cell source of cartilage regeneration but has several issues to overcome such as limited cell proliferation and heterogeneity of cartilage regeneration ability. Previous reports demonstrated that basic fibroblast growth factor (bFGF) can promote proliferation and cartilage differentiation potential of MSCs in vitro, although no reports show its beneficial effect in vivo. The purpose of this study is to investigate the promoting effect of bFGF on cartilage regeneration using human SMSC in vivo. SMSCs were cultured with or without bFGF in a growth medium, and 2 × 105 cells were aggregated to form a synovial pellet. Synovial pellets were implanted into osteochondral defects induced in the femoral trochlea of severe combined immunodeficient mice, and histological evaluation was performed after eight weeks. The presence of implanted SMSCs was confirmed by the observation of human vimentin immunostaining-positive cells. Interestingly, broad lacunae structures and cartilage substrate stained by Safranin-O were observed only in the bFGF (+) group. The bFGF (+) group had significantly higher O’Driscoll scores in the cartilage repair than the bFGF (−) group. The addition of bFGF to SMSC growth culture may be a useful treatment option to promote cartilage regeneration in vivo.


Introduction
Articular cartilage has a poor self-healing ability and the conventional surgical treatments and cell therapies for cartilage injuries, such as bone marrow stimulation, autologous osteochondral transplantation, and autologous chondrocyte implantation (ACI), have not yielded full recovery of normal cartilage [1]. Furthermore, ACI requires two-step surgery and the sacrifice of normal cartilage tissue to supply new cartilaginous tissue to the defect. Therefore, cartilage regeneration using mesenchymal stem cells (MSCs) has been widely researched [2,3]. Synovial MSCs (SMSCs) are an attractive cell source for cartilage regeneration because they are easily collected as surplus tissue during joint surgery and show favorable chondrogenesis and proliferation potential both in vitro [4,5] and in vivo [6,7]. During the development of commercialization of an allogenic SMSC bank for cartilage repair, we have reported that SMSCs obtained even from young patients showed a large variation in their proliferation and cartilage differentiation potentials [8]. In fact, our human clinical trial revealed that autologous SMSC implantation for the repair of a cartilage defect failed to recover full articular cartilage [9]. Therefore, various pretreatment methods to enhance the cartilage differentiation potential of MSCs have been reported [10][11][12]; we showed that a low oxygen tension prevented cellular senescence and promoted the proliferative and chondrogenic differentiation capacity of human SMSCs in vitro [13]. However, none of these reports succeed to demonstrate the positive effects in vivo. The basic fibroblast growth factor (bFGF) has been widely reported to enhance the potential of proliferation and cartilaginous differentiation of MSCs in vitro [14][15][16]. Studies of embryonic, neural, and bone marrow stem cells reported that bFGF regulates undifferentiated state and multipotent differentiation of stem cells through the fibroblast growth factor receptor 3 (FGFR3)-mediated signaling [14,17,18]. FGFR3 is known as one of the markers of mesenchymal pre-cartilaginous stem cells and was detected at the margins of the cartilage nodules of synovial chondromatosis and SMSCs [19][20][21]. However, no positive effects have been documented in vivo. We hypothesized that SMSCs pretreated with bFGF may demonstrate enhanced cartilaginous differentiation in vivo. The purpose of this study was to investigate whether bFGF can promote cartilage repair using human SMSCs in vivo.

Expression of FGFR3 Protein
To analyze the expression of FGFR3 in the SMSCs from different patients, Western blotting was conducted using anti-FGFR3 antibody. The FGFR3 protein expression was confirmed in all of the patients' SMSCs by Western blotting (Figure 2). (c) Cell counts of SMSCs in the bFGF (+) group at each passage relative to those of the bFGF (−) group (patients #1 to #5). ** p < 0.01. bFGF, basic fibroblast growth factor; SMSCs, synovial mesenchymal stem cells.

Expression of FGFR3 Protein
To analyze the expression of FGFR3 in the SMSCs from different patients, Western blotting was conducted using anti-FGFR3 antibody. The FGFR3 protein expression was confirmed in all of the patients' SMSCs by Western blotting (Figure 2).

Expression of Cell Surface Markers
To evaluate the effects of bFGF on the expression of cell surface markers in SMSCs, fluorescence-activated cell sorting (FACS) was conducted for four mesenchymal surface markers (CD44, CD73, CD90, and CD105), two mesenchymal negative markers (CD11b, CD271), and one chondrocyte surface marker (CD151) (Figure 3). Almost 100% of the SMSCs expressed all four mesenchymal surface markers and one chondrocyte marker, and almost none of SMSCs expressed mesenchymal negative markers. Moreover, their expression levels were not affected by bFGF.  (c) Cell counts of SMSCs in the bFGF (+) group at each passage relative to those of the bFGF (−) group (patients #1 to #5). ** p < 0.01. bFGF, basic fibroblast growth factor; SMSCs, synovial mesenchymal stem cells.

Expression of FGFR3 Protein
To analyze the expression of FGFR3 in the SMSCs from different patients, Western blotting was conducted using anti-FGFR3 antibody. The FGFR3 protein expression was confirmed in all of the patients' SMSCs by Western blotting (Figure 2).

Expression of Cell Surface Markers
To evaluate the effects of bFGF on the expression of cell surface markers in SMSCs, fluorescence-activated cell sorting (FACS) was conducted for four mesenchymal surface markers (CD44, CD73, CD90, and CD105), two mesenchymal negative markers (CD11b, CD271), and one chondrocyte surface marker (CD151) (Figure 3). Almost 100% of the SMSCs expressed all four mesenchymal surface markers and one chondrocyte marker, and almost none of SMSCs expressed mesenchymal negative markers. Moreover, their expression levels were not affected by bFGF.

Expression of Cell Surface Markers
To evaluate the effects of bFGF on the expression of cell surface markers in SMSCs, fluorescence-activated cell sorting (FACS) was conducted for four mesenchymal surface markers (CD44, CD73, CD90, and CD105), two mesenchymal negative markers (CD11b, CD271), and one chondrocyte surface marker (CD151) (Figure 3). Almost 100% of the SMSCs expressed all four mesenchymal surface markers and one chondrocyte marker, and almost none of SMSCs expressed mesenchymal negative markers. Moreover, their expression levels were not affected by bFGF.

The Effect of bFGF on the Chondrogenesis of SMSCs in a Three-Dimensional (3D) Chondrogenic Culture
To assess the effect of the addition of bFGF in the cell-proliferation phase on cartilaginous tissue formation, a 3D pellet culture was established. After four weeks of chondrogenic induction, histology, the expression of cartilage differentiation-related genes, and extracellular matrix (ECM) production, were evaluated. Synovial pellets of the bFGF (+) group grew larger and were more strongly stained with Safranin-O (Saf-O) and immunostained for collagen type II (COL2) than those of the bFGF (−) group (n = 5 per group) (   To assess the effect of the addition of bFGF in the cell-proliferation phase on cartilaginous tissue formation, a 3D pellet culture was established. After four weeks of chondrogenic induction, histology, the expression of cartilage differentiation-related genes, and extracellular matrix (ECM) production, were evaluated. Synovial pellets of the bFGF (+) group grew larger and were more strongly stained with Safranin-O (Saf-O) and immunostained for collagen type II (COL2) than those of the bFGF (−) group (n = 5 per group) (Figure 4a). A similar tendency was observed in all patients. Moreover, pellets of the bFGF (+) group showed hyaline cartilage-like differentiation with chondrocyte-like round-shaped cells in the lacuna, abundant collagen stained with Sirius Red and abundant ECM stained with Saf-O and Alcian Blue (Figure 4b).

In vivo Osteochondral Repair in the Mouse Model Using Pellets Generated From the SMSCs From the bFGF (−) and bFGF (+) Groups
To assess the effect of the pretreatment with bFGF on cartilaginous regeneration of SMSC in vivo, synovial pellets were implanted into the osteochondral defects created in the knee of SCID mice. Eight weeks after implantation, no grossly visible synovitis or inflammation (Figure 6a) was observed in mouse knees in all groups, and no apparent infiltration of inflammatory cells into the repair cartilage was detected by hematoxylin and eosin (HE) staining in the two groups ( Figure 6b). In both pellet-treated groups, the implanted cell aggregates were retained, and a little bony repair and a sclerotic rim formed around the implanted pellets were observed in the defects of the subchondral region. , and ACAN normalized to GAPDH (10 pellets from patients #3 to #5). ** p < 0.01; *** p < 0.001. bFGF, basic fibroblast growth factor; SMSCs, synovial mesenchymal stem cells; 3D, three-dimensional; CSA, cross-section area; COL2, collagen type II; sGAG, sulfated glycosaminoglycan; dsDNA, double-stranded DNA; COL2A1, collagen type II alpha 1; COL10A1, collagen type X alpha 1; SOX9, sex-determining region Y-box 9; ACAN, aggrecan; GAPDH, glyceraldehyde 3-phosphate dehydrogenase.

In vivo Osteochondral Repair in the Mouse Model Using Pellets Generated From the SMSCs From the bFGF (−) and bFGF (+) Groups
To assess the effect of the pretreatment with bFGF on cartilaginous regeneration of SMSC in vivo, synovial pellets were implanted into the osteochondral defects created in the knee of SCID mice. Eight weeks after implantation, no grossly visible synovitis or inflammation (Figure 6a) was observed in mouse knees in all groups, and no apparent infiltration of inflammatory cells into the repair cartilage was detected by hematoxylin and eosin (HE) staining in the two groups ( Figure 6b). In both pellet-treated groups, the implanted cell aggregates were retained, and a little bony repair and a sclerotic rim formed around the implanted pellets were observed in the defects of the subchondral region.
However, the defects of the subchondral bone region were filled with repaired bone in the defect-only group. Human-vimentin-positive cells were found in the pellet implantation site in both the bFGF (−) and bFGF (+) groups, suggesting the existence of implanted human SMSCs (Figure 6c).

Discussion
Fibroblast growth factors (FGFs) are expressed widely in various tissues and display various biological properties [21]. FGFs control the migration, proliferation, differentiation, and survival of various cells and affect the expression of other factors involved in the regenerative response [22]. The human-mouse FGF family comprises 22 members [23], and the bFGF (also known as FGF2) is a member of the FGF family. Most FGFs bind to heparinsulfate proteoglycan for signaling through the fibroblast growth factor receptor (FGFR) [24] and FGFR1-4 has been identified in both humans and mice [21]. As mentioned, FGFR3 is one of the markers of mesenchymal pre-cartilaginous stem cells, and bFGF regulates undifferentiated state and multipotent differentiation of stem cells through FGFR3-mediated signaling [14,17,18]. In the present study, the FGFR3 protein was expressed in the SMSCs of all donors assessed by Western blotting, and the addition of bFGF to the growth medium changed the morphology of SMSCs and promoted its proliferation rate without altering the expression of mesenchymal surface markers (CD44, CD73, CD90, and CD105), and the chondrocyte marker (CD151). These results imply bFGF/FGFR3 could promote chondrogenic differentiation of MSCs without changing their population. SOX9 is known as a master chondrogenic factor [25] and involved in the early phase of chondrogenesis and chondrocyte proliferation [26]. bFGF upregulates SOX9 gene expression for at least first 24 h [27] and induces chondrocyte proliferation [28]. In the present study, however, SOX9 was not significantly affected by bFGF administration. This may be due to the timing of evaluation, because SOX9 gene expression was assessed at 28 days after bFGF stimulation, which may be too late to detect SOX9 upregulation by bFGF. On the other hand, bFGF controlled the maintenance of stemness in MSCs [29,30], and suppressed the cellular senescence of MSCs by downregulating transforming growth factor β2 (TGF-β2) [31]. Cellular senescence downregulates the therapeutic potential of human MSCs by reducing their migratory, homing, and immunoregulatory abilities in vivo [32,33]. The binding interaction between chondrocytes and the ECM is critical for maintaining chondrocyte activity [34]. Altogether, these findings suggest that suppressed cellular senescence and enhanced ECM production by bFGF may lead to enhanced cartilage regeneration in vivo.
Regarding the mechanism underlying the in vivo tissue repair provided by MSCs, studies have demonstrated that transplanted MSCs differentiate directly into repaired tissue [35,36]. However, other reports concluded that MSCs lead to enhanced reparative response via the secretion of growth factors, cell-cell interactions, and the release of extracellular vesicles, which contain peptides/proteins, mRNA, and microRNAs [37][38][39]. Likewise, it is unclear whether MSCs differentiate into chondrocytes or promote host-side cartilage repair [40]. Ozeki et al. reported that synovial MSCs from a green fluorescent protein-expressing rat injected into the rat's knee joint were detectable three weeks after injection [41]. In contrast, 1,1 -dioctadecyl-3,3,3 ,3 -tetramethylindocarbocyanine perchloratelabeled pig synovial MSCs implanted into osteochondral defects of the knee were detectable at one week but could not be found at one month, after implantation [42]. In the present study, human synovial MSCs were clearly detectable 8 weeks after the operation, as assessed by human vimentin immunostaining. Moreover, synovial pellets of the bFGF (+) group exhibited a higher cartilage differentiation potential in vivo than those of the bFGF (−) group. It has been previously reported that chondrocytes overexpressing bFGF are effective in repairing osteochondral defects [43,44]. However, to the best of our knowledge, there are no reports demonstrating that adding bFGF in the culture medium enhanced the cartilage tissue repair ability of SMSCs, which is by far a simple method. Regarding tissue engineering, we developed a scaffold-free three-dimensional tissue-engineered construct (TEC) with SMSCs, which produces rich ECM [45], and reported its cartilage repair efficacy in nude rats [8] and humans [9]. In the present study, instead of TEC, we used simple synovial pellet to simplify the methods. However, unlike our previous reports using TEC [7,8], the cartilage repair ability using simple synovial pellet without bFGF was generally low. In addition, an osteosclerotic image at the boundary with the subchondral bone as well as bone remodeling delay was observed in both bFGF(−) and bFGF(+) groups. These results suggest that synovial pellet transplantation may have prevented the repair of the host-side subchondral bone in the process of producing ECM and differentiating into cartilage tissue, and further improvement in transplantation methods may be required.
This study had several limitations. First, because of the necessity of using human SMSCs, we had to employ an in vivo xenograft model using a severe combined immunodeficient (scid) mouse, in which immunological reactions may be excessively suppressed. Second, the establishment of a full cartilage defect model using a large animal model and a longer follow-up is required to evaluate the mechanical loading and long-term maintenance of the regenerated cartilage. However, to the best of our knowledge, the results of this study demonstrated for the first time that addition of bFGF to SMSC growth culture may be a simple and useful treatment option to promote cartilage regeneration both in in vitro and in vivo.

Materials and Methods
All procedures were performed in accordance with the Declaration of Helsinki. All experiments carried out were approved by the Ethical Review Board of the Osaka University Hospital (number 15409-2, approval date April 12, 2016) and the Animal Experiment Regulations of our institution. Written informed consent was obtained from all patients.

Patient Demographics
The human synovium was obtained from five patients (one male and four female patients; average age, 17.4 years; range, 13-22 years) at the time of arthroscopic surgery of the knee joint. Among them, four patients underwent anterior cruciate ligament reconstruction with arthroscopy and one patient underwent a second look with arthroscopy after internal fixation of osteochondritis dissecans. The average waiting period until surgery was 13.2 weeks (Table 2).

Isolation and Culture of Human SMSCs
The harvested synovium (average wet weight, 483 mg; range, 400-590 mg) was stored in the 0.9% sodium chloride irrigation (Baxter, Deerfield, IL, USA) at room temperature and SMSCs were isolated within at least 3 h (Figure 8a). The synovium was sheared into small fragments and digested with 0.25% trypsin/EDTA (Gibco, Carlsbad, CA, USA) for 30 min, followed by extraction of floating adipose tissues after centrifugation [46]. Non-floating tissues were resuspended in Dulbecco's modified Eagle medium (DMEM) (Gibco) containing 10% fetal bovine serum (FBS) (Sigma-Aldrich, St. Louis, MO, USA) and supplemented with 400 U/mL of collagenase type A (Worthington, Lakewood, NJ, USA), followed by incubation at 37 • C. Three hours later, dissociated cells were collected, washed by centrifugation, and cultured as described below. Isolated SMSCs were separated into two groups: the bFGF (−) group and the bFGF (+) group. SMSCs of the bFGF (−) group were cultured in high-glucose DMEM containing 10% FBS and 1% antibiotic-antimitotic (Sigma-Aldrich) growth medium (GM), and SMSCs of the bFGF (+) group were cultured in GM with 5 ng/mL bFGF (Wako, Osaka, Japan) at 37 • C with humidified 5% CO 2 (Figure 8b). Then, 5 ng/mL bFGF addition was adopted with reference to previous reports [16,19,47]. SMSCs were expanded as described previously and used for experiments (analyses of cell morphology, growth profile, and FACS) at passage 3 [46]. exposed via a medial parapatellar incision with lateral patellar dislocation, and an osteochondral defect of 1.0 mm diameter and depth was created in each knee using a micro drill (Kiso power tool, Osaka, Japan, Cat# 28854, 28855), and a microscope. A line was marked 1.0 mm from the tip of the micro drill. Subsequently, 12 knees were left empty as a defect-only group, and 26 knees were implanted with synovial pellets using a microscope. Pellets from the bFGF (−) and bFGF (+) groups were implanted randomly, and one pellet was implanted in each defect.
Eight weeks after surgery, the scid mice were sacrificed and their knees were harvested for analyses (Figure 8e).

Cell-Proliferation Assay
The cell-proliferation capacity of SMSCs from patients #1 to #5 was assessed from passages 1 to 6 by evaluating the population doubling level (PDL). For this, 3 × 10 5 cells/dish were seeded onto 150-mm dishes and cultured to reach approximately 90% confluence, followed by passaging after cell counting. The PDL was calculated using the following formula: PDL = log(N/N0)/log2, where N0 is the number of plated cells and N is the number of harvested cells at the time of passage. The cumulative PDL was the total PDL = Σ(PDL)n, where n is the passage number [48].

Implantation of Synovial Pellets onto Osteochondral Defects
For implantation, 2 × 10 5 cells harvested from passage 3 were centrifuged in 96 deepwell polypropylene plates and cultured in GM (Figure 8c), and surgery was performed on the following day. Ten-week-old male scid mice (C.B-17/Icr-scid/scidJcl; CLEA Japan, Fujinomiya, Japan) were anesthetized with an intraperitoneal injection of 0.3 mg/kg medetomidine (Nippon Zenyaku Kogyo Co., Ltd., Fukushima, Tokyo, Japan), 4.0 mg/kg midazolam (Astellas Pharma, Inc., Tokyo, Japan), and 5.0 mg/kg butorphanol (Meiji Seika Pharma Co., Ltd., Tokyo, Japan) [52]. In both knees, the femoral trochlear grooves were exposed via a medial parapatellar incision with lateral patellar dislocation, and an osteochondral defect of 1.0 mm diameter and depth was created in each knee using a micro drill (Kiso power tool, Osaka, Japan, Cat# 28854, 28855), and a microscope. A line was marked 1.0 mm from the tip of the micro drill. Subsequently, 12 knees were left empty as a defect-only group, and 26 knees were implanted with synovial pellets using a microscope. Pellets from the bFGF (−) and bFGF (+) groups were implanted randomly, and one pellet was implanted in each defect.
Eight weeks after surgery, the scid mice were sacrificed and their knees were harvested for analyses (Figure 8e).

Histology and Immunohistochemistry
All samples (pellets and distal femurs) were fixed in 4% paraformaldehyde and embedded in paraffin wax, followed by dehydration with an ethanol series and clearance with xylene. To evaluate osteochondral tissues, the dissected femoral ends were decalcified with 10% ethylenediaminetetraacetic acid (pH 7.4) before paraffin embedding. The samples were cut into 5-µm-thick sections and used for HE staining, toluidine blue staining, Saf-O staining, and immunohistochemistry, as described previously [51,53]. The following antibodies were used: anti-COL2 (Kyowa Pharma Chemical, Toyama, Japan, Cat# F-57, 1:500 dilution), anti-COL10 (Thermo Fisher Scientific, Cat# 14-9771-82, 1:500 dilution), and anti-human vimentin (Abcam, Cambridge, UK, Cat# ab16700, 1:100 dilution). The CSA of pellets and COL2 staining were measured automatically using an Aperio Image Scope (Leica Biosystems, Wetzlar, Germany). The sections stained with HE and Saf-O were used for histological evaluation using the O'Driscoll scoring system for cartilage repair, as described previously [54].

Statistical Analysis
All data were expressed as the mean ± standard deviation. Differences between two groups were assessed using the Mann-Whitney U test. Differences between three groups were assessed using the Kruskal-Wallis test. Significance was set at p < 0.05. Statistical analyses were performed using EZR (Saitama Medical Center, Jichi Medical University, Saitama, Japan) [55].  Data Availability Statement: Data available from the corresponding author on reasonable request.

Acknowledgments:
We thank Akira Tsujii, Yukiko Eguchi, Fumiko Hirayama, and Mari Shinkawa for excellent assistance. We thank all the members of Yoshikawa's laboratory for the helpful discussion and comments.

Conflicts of Interest:
The authors declare no conflict of interest.