Isolation and Characterization of Synovial Mesenchymal Stem Cell Derived from Hip Joints: A Comparative Analysis with a Matched Control Knee Group

Purpose. To determine the characteristics of MSCs from hip and compare them to MSCs from knee. Methods. Synovial tissues were obtained from both the knee and the hip joints in 8 patients who underwent both hip and knee arthroscopies on the same day. MSCs were isolated from the knee and hip synovial samples. The capacities of MSCs were compared between both groups. Results. The number of cells per unit weight at passage 0 of synovium from the knee was significantly higher than that from the hip (P < 0.05). While it was possible to observe the growth of colonies in all the knee synovial fluid samples, it was impossible to culture cells from any of the hip samples. In adipogenesis experiments, the frequency of Oil Red-O-positive colonies and the gene expression of adipsin were significantly higher in knee than in hip. In osteogenesis experiments, the expression of COL1A1 and ALPP was significantly less in the knee synovium than in the hip synovium. Conclusions. MSCs obtained from hip joint have self-renewal and multilineage differentiation potentials. However, in matched donors, adipogenesis and osteogenesis potentials of MSCs from the knees are superior to those from the hips. Knee synovium may be a better source of MSC for potential use in hip diseases.


Introduction
An acetabular labral tear is one of the most common sources of hip pain and disability and one of the most frequent indications for hip arthroscopic surgeries. A concomitant labral tear with underlying femoroacetabular impingement (FAI) and/or dysplasia are risk factors for future cartilage damage and hip osteoarthritis [1].
Several standard procedures, including debridement and microfracture, have been routinely utilized to address chondral damage of the hip joint [2,3]. In fact, there are studies that demonstrate favorable short to medium outcomes following microfracture treatment of localized cartilage damage in patients with FAI [4][5][6]. Other promising procedures such as osteochondral transplantation and autologous chondrocyte implantation (ACI) have been attempted to preserve cartilage for FAI patients [7]. However, studies on these newer procedures report consistently poor outcomes with high rates of conversion to total hip arthroplasty following hip arthroscopy for FAI patients with early to moderate osteoarthritis at the time of surgery [6,[8][9][10]. Thus, patients with severe cartilage damage are not good candidates for hip preservation surgery due to the high risk of failure.
Minimally invasive stem cell therapy may be a desirable strategy to treat cartilage lesions in the hip. Mesenchymal stem cells (MSCs) are incredibly unique candidates for 2 Stem Cells International use of cell therapy as they can be isolated from various tissue reservoirs such as bone marrow, adipose tissue, bone tissues, and synovial tissues [11,12]. Previous in vitro studies demonstrated that MSCs can be isolated from synovial fluid in the knee joints [13]. Synovial MSCs derived from synovial tissues are superior to these other sources such as bone marrow and adipose tissues [14]. In addition, a recent in vivo study demonstrated that MSCs derived from synovial tissue have greater cartilage regenerating potential through comparing histological samples following knee injection in porcine animal models [15]. Sekiya et al. were able to translate this finding clinically and reported good outcomes following the use of synovial MSCs during arthroscopy for patients with cartilaginous knee injuries [16][17][18]. Much of the available literature is on MSCs derived from knee synovial tissues with a dearth of knowledge regarding MSCs derived from hip synovial tissues. Therefore, it is essential to evaluate the proliferation and differentiation potentials of MSCs derived from hip joints for future applications using MSCs therapy in the hip. This source of cells could potentially be beneficial for the treatment of hip disease such as focal chondral lesions and early OA.
The purpose of this study was (i) to isolate and define the characteristics of MSCs from synovial fluid and synovial tissues of the hip joint and (ii) to clarify whether there are any differences between synovial MSCs from the knee and the hip joint in the same donor. We hypothesize that MSCs derived from synovial fluid and synovial tissues of the hip joint can be isolated with the same proliferation and differentiation potential as those from the knee joint.

Patient Selection, Tissue Collection, and Isolation of Cells.
Approval for the study was granted through the institutional review board. 423 hip arthroscopic surgeries were performed by a single surgeon between 2012 and 2014. Of 423 patients, 8 were eligible for this study as they suffered from both hip and knee intra-articular pathologies. Synovial samples were obtained from 2 male and 6 female patients (mean age ± SD; 47.9 ± 14.0 (range 25-64) years) during knee and hip arthroscopic surgeries. There were 3 premenopausal females and 3 postmenopausal females. There was no significant difference between the durations from onset of symptoms to date of surgery in these 8 patients. Furthermore, the preoperative osteoarthritis Kellgren-Lawrence grade on radiological classification between both joints for these patients was comparable (Table 1).
Arthroscopic surgeries for knee and hip pathologies in each eligible patient were performed on the same day. Human synovial fluid and synovium of the knee and the hip were obtained during arthroscopic surgery under general anesthesia. To minimize variation in sampling, specimens were harvested from the locations near the tear of meniscus in the knee and the torn labrum in the hip (Figure 2(a)). In Figure 2(a), when we harvested hip synovium, we released tissue junction between capsular ligament (iliofemoral ligament) and synovium and used a rongeur turning back on capsule to avoid the contamination of the capsule. Synovial fluid was obtained from the knee and the hip joints after injection of saline solution (50 mL). Synovium of the knee was obtained near the peripheral site at 5 mm distance from the injury of meniscus with a rongeur under arthroscopic observation. Synovium in the hip was also obtained from the peripheral area at 5 mm distance adjacent to the labral tear by using the same procedure. Arthroscopic findings of meniscus, labrum, and cartilage at the surgery and surgical procedure at both joints are shown in Table 1.
Both synovial samples were digested in 3 mg/mL collagenase from Type V (Sigma-Aldrich, St. Louis, Missouri) in -modified Eagle's medium ( -MEM, Invitrogen, Carlsbad, CA) at 37 ∘ C. After 2 hours, the digested nucleated cells of synovium were plated at 10 3 , 10 4 , and 10 5 cells per 60 cm 2 culture dish (Nalge Nunc International, Rochester, NY), with 6 dishes created for each cell concentration. Cells were seeded in complete culture medium containing -MEM and 20% fetal bovine serum (FBS). A high concentration of FBS was used to induce the highest number of colony-forming units after cell culture derived from hip synovial fluid among 6 lots of FBS. The cells were incubated at 37 ∘ C with 5% humidified CO 2 . After 3 to 4 days, the medium was changed to remove nonadherent cells. The adherent cells were then cultured for 14 days as passage 0 without refeeding. After 14 days, three dishes for each concentration were stained with 0.5% crystal violet. The optimal initial cell density was determined by (1) the resultant colony size being unaffected by contact inhibition and (2) a maximum in number of colonies being reached. We then harvested the cells plated at optimal densities from the remaining 3 dishes with Trypsin-EDTA (Invitrogen) and determined the number of cells at passage 0. The passage 0 cells were replated at 50 cells/cm 2 in a 150 cm 2 dish and cultured for 21 days. The cultured cells were cryopreserved by resuspension at a concentration of less than 10 6 cells/mL in 10% fetal bovine serum. One-milliliter aliquots were slowly frozen to avoid cell damage and cryopreserved in liquid nitrogen (passage 1). To expand the cells, we thawed a frozen vial of the cells, plated the cells in 60 cm 2 culture dishes, and incubated them for 4 days in a recovery plate. These cells (passage 2) were used for further analyses as previously described [19].

Estimation of the Yield Obtained.
To investigate proliferation potential of synovial MSCs from the knee and the hip joints, we estimated the number of live cells at passage 0 gained from whole volume of each tissue and compared it to the final count of live cells from our cultures.

Histological Analysis.
Synovial tissues obtained from the knee and the hip joints were fixed in a solution of 4% paraformaldehyde (PFA) with 0.1 mol/L phosphate buffer (PB) (pH 7.2), dehydrated through a graded series of alcohol, and embedded in paraffin. 5 m slice sections were prepared on microtome. These sections were stained with hematoxylin and eosin.

In Vitro Expandability and Cell Viability Assay.
To examine the in vitro expandability of cells from each type of mesenchymal tissue, passage-1 cells were replated at 50 cells/cm 2 in 150 cm 2 culture dishes every 14 days until their growth potential was exhausted [14]. To determine the viabilities of cells in each passage, we analyzed quantitative evaluation on 3 donors with use of LIVE/DEAD Viability/Cytotoxicity Kit for mammalian cells (Invitrogen). In addition, a colorimetric assay for cell viability was conducted by plating passages 4 and 7 cells at 5000 cells in a 24-well multidish (Nalge Nunc International). These cells were cultured with 500 L of complete medium for 7 days. After incubation of the cells, 50 L of Cell Counting Kit-8 (highly water-soluble tetrazolium salt, WST-8, Dojindo, Kumamoto, Japan) was added to each well, and the dish was incubated for 1 hour. Cells were counted with a spectrophotometer multiplate reader 450 nm (SH1000, Corona Electric, Ibaraki, Japan). Cell viability was evaluated and expressed as a percentage based on the viable cell count of the synovium in knee (100%). At last, we analyzed the correlation between the fold increase and cell viability on each of the samples.

Osteogenesis in a Colony-Forming Assay.
One hundred passage-2 cells were plated in 60 cm 2 dishes and cultured for 14 days. The medium was then switched to osteogenic differentiation medium consisting of con supplemented with 10% FBS, 10 −9 M dexamethasone, 20 mM -glycerophosphate (WAKO), and 50 g/mL ascorbate-2-phosphate for an additional 21 days. After 4% PFA fixation and positive colonies were counted, mineralized bone nodule formation was assessed by double labeling for mineral (von Kossa stain) and alkaline phosphate (ALP). The same osteogenic cultures were subsequently stained with crystal violet and the total cell colonies were counted. Colonies that were <2 mm in diameter and appeared yellow in color were discounted [14].

Immunohistochemistry.
Four-micrometer-thick sections obtained from the tissue specimens were deparaffinized and incubated with primary anti-human collagen II antibody (ab34712, Abcam, Cambridge, UK) for 30 min at 4 ∘ C. The sections were rinsed again with PBS repeatedly and then incubated with immune-peroxidase polymer secondary antibody (Histofine, Nichirei Biosciences, Tokyo, Japan) for 30 min. Antibodies were visualized by treating the sections with diaminobenzidine tetrahydrochloride (Chem-Mate DAB+ Chromogen, DakoCytomation) for 5 min. To assess the specificity of immunohistochemistry of COLII, the sections of human chondrosarcoma were stained as described above as our positive control.

Quantitative Real-Time RT-PCR.
Cells that entered adipogenesis or osteogenesis were trypsinized, and total RNA was obtained with RNeasy mini plus kit (Qiagen) according to the manufacturer's protocol. The chondrogenic pellet was homogenized using BioMasher (Nippi, Tokyo, Japan), and total RNA was prepared from a single pellet using RNeasy micro plus kit (Qiagen, Hilden, Germany). After total RNA was reverse-transcribed, cDNA was assayed three times with 10 TaqMan   normalization of gene expression (data not shown). StepOne-Plus6 Real-Time RT-PCR System was used for the quantification and analysis was performed using StepOnePlus6 software (version 2.0, Applied Biosystems).

Statistical Analysis.
To assess the difference in isolation status and the character of cells from the synovium of knee and hip, we used Mann-Whitney test. Statistical analyses were performed using SPSS (version 13, SPSS Inc., Chicago, IL) software package. The level of significance was set at a probability value of <0.05.  Table 2.

Macroscopic and Histological
The nucleated cell number per milligram of the synovium from knee was significantly higher than that of the hip (knee: However, it was difficult to assess whether there were significant differences because only two male patients were included in the study. In order to gain a maximum number of colonies per amount of nucleated cells, we examined the effect of plating density on nucleated cells from each type of mesenchymal tissue from the 8 donors. From synovium of the knee and the hip, large single cell-derived colonies occurred when nucleated cells were plated at 10 4 cells/60 cm 2 dish. The optimal initial cell density was determined as 10 4 /60 cm 2 dish (Figure 2(f)). There was no significant difference in the number of colony-forming units per 10 4 nucleated cells between knee and hip ( = 0.33) (Figure 1 and Table 2). The cells from both types of synovium seemed to be similar in fibroblastic cell morphology (Figure 2(g)).
There was no significant difference of CFU-f regarding gender difference in each of the joints.    hip synovial fluid (Table 3 and Figure 2(h)). Thus, analysis to compare number of colonies between the synovial fluids of the knee and the hip was not performed. The number of colony-forming fibroblasts was 3.3 ± 4.9 in the knee joints.

In Vitro Expandability and Cell Viability Assay.
Expandability was lost after eight passages for both types of synovium-derived cells (Figure 3). Using a LIVE/DEAD Viability/Cytotoxicity Kit, both types of passages 1 to 3 cells were determined to be alive. In passage 4, cells derived from the knee synovium were thriving but cells derived from the hip synovium were losing viability (Figure 4(a)). Using a Cell Counting Kit-8 assay, viability of cells derived from the knee synovium was significantly higher than that of cells derived from the hip synovium (knee versus hip, 100% versus 88.4%, < 0.01, Mann-Whitney U test) (Figure 4(b)). On the other hand, the half of both cells was dead at passage 7. At passage 7, there was no significant difference of cell viability between knee and hip with both cell lines nonviable. In addition, there was a strong correlation between the fold increase and cell viability ( 2 = 0.71) (Figure 4(c)).

Adipogenesis.
In an adipogenesis differentiation assay, all samples generated colonies of adipocytes, which stained with Oil Red-O ( Figure 5(a)). However, the rate of Oil Red-O-positive colonies was significantly higher in cells derived from the knee as compared to those from the hip ( Figure 5(b); < 0.05, Mann-Whitney U test). In addition, real-time quantitative PCR analysis for adipogenic gene expression was performed ( = 3). As shown in Figure 5(c), the gene expression of adipsin in the knee group was significantly higher than those in the hip ( < 0.05, Mann-Whitney test). There were no significant differences in gene expression of PPARG and LPL between both joints ( = 0.26 and 0.26, Mann-Whitney test).

Osteogenesis.
In an osteogenesis differentiation assay, cells derived from the knee synovium generated colonies of osteogenic cells that stained with von Kossa/ALP double staining (Figure 6(a)). Rates of von Kossa-and ALP-positive colonies in relation to the total number of colonies between cells derived from the knee and the hip joints were found to be not statistically different between the knee and the hip (Figure 6(b)) ( = 0.20, Mann-Whitney U test). According to real-time PCR assay, the levels of gene expressions of COL1 and ALP in cells from the hip were significantly higher than those in cells from the knee (3 donors, Figure 6(c); < 0.01 (COL1A1), = 0.04 (ALP)). In contrast, the gene expressions of RUNX2 and BGLAP genes in cells from the knee were significantly higher than those in cells from the hip (3 donors, Figure 6(c); < 0.01 (RUNX2), < 0.01 (BGLAP), Mann-Whitney test). These findings indicate that osteogenic potential of MSCs from the hip joint is equivalent to that from the knee joint, but osteogenic rate of MSCs from the knee joint is greater than that from the hip joint [21].

Chondrogenesis.
In chondrogenesis differentiation assays, most of the pellets from each tissue were >0.5 mm in diameter. The pellets of both samples had positive Toluidine Blue stained cartilage matrix. There was no significant difference of the mean diameter of pellets between both pellets (knee: 825 m versus hip: 800 m, = 0.82, Mann-Whitney test). Immunoreactivity for type II was found in both pellets (Figure 7(a)) but immunoreactivity for type X collagen was not. The real-time PCR study demonstrated that the gene expression of Col2A1 in pellets derived from the knee was significantly higher than that in pellets derived from the hip (Figure 7(b); < 0.01, Mann-Whitney test). There were no significant differences of the expressions of COL10A1 and SOX9 between both pellets (Figure 7(b)). These findings suggest that the chondrogenic potential of MSCs from the hips is nearly comparable to that from the knee joint despite the expression of Col2A1 from the knee MSCs being superior to that from the hip MSCs.

Discussion
This study has four major findings. First, we determined that adherent cells derived from the synovium of the hip joints can be isolated with consistent proliferation and multilineage differentiation potentials. Second, we found that colonyforming units, fibroblastic morphology, and epitope profile of MSCs derived from synovium of the hip joints were similar to MSCs derived from synovium of the knee joints. Third, there was a difference in the number of nucleated cells derived from synovium per milligram, colony-forming units derived from synovial fluid, proliferation rate, adipogenesis, osteogenesis, and the expression of COL2A1 during chondrogenesis in MSCs synovium from the knee joints compared to the hip joints. Fourth, the yield obtained from synovium of the knee joints was slightly higher but not more significant than the hip joints.
One of the strengths of this study is that we compared the proliferation and differentiation potentials between synovial MSCs of two joints from the same donor. Previous studies demonstrated that the condition of the donors could potentially affect the ability of MSCs to function and suggested that MSCs differ in property and vary from donors to donors [24][25][26]. By providing same donor samples, we are avoiding this potential bias. Siegel et al. demonstrated that age and gender affect the growth potential of MSCs derived from human bone marrow [26]. Lee et al. demonstrated that estradiol affected MSCs' potential and proliferation rate in a porcine animal model [27]. They concluded that MSCs of females were affected by menopause. Matsukura et al. demonstrated that a longer period of time between onset of pain and surgery in patients with meniscus injury increased the amount of MSCs derived from human synovial fluid [24]. Sekiya et al. also demonstrated that MSCs derived from human synovium in patients with severe osteoarthritis have greater potential of proliferation compared to healthy patients [25]. However, they did not find any significant difference when comparing onset of pain to surgery or severity of osteoarthritis (OA).
In this study, we isolated adherent cells derived from synovium of the hip and the knee joints. Our findings demonstrated that these synovial cells from both the knee and the hip joints have proliferation and multilineage differentiation potentials. MSCs from both joints had a morphologically spindle shape, similar to that of fibroblasts, with functional capacities to proliferate and differentiate into multilineage cells. In addition, cells from both joints expressed similarly surface markers. Sakaguchi et al. and Segawa et al. demonstrated that cells derived from human synovium and bone marrow were negative for CD34 and CD45 and positive for CD90, CD105, and CD44 [14,20]. Similarly, our findings reveal that expressions of both cells were negative for CD34, CD45, CD31, CD117, NGFR, and VEGFR2 and positive for CD44, CD90, CD105, and CD147. These findings indicate that both knee and hip synovial derived MSCs expressed similar cell surface epitopes to those from other cell sources including bone marrow and adipose tissues. International Society for Cellular Therapy position statement defined that MSCs are as follows: (1) adherent cells to plastic, (2) specific surface marker expression, and (3) multipotent differentiation potential. In this study, we were able to define MSCs from both the hip and the knee joints according to the criteria stated above.
In this study, we investigated the yield obtained and expandability as part of the proliferation potential of MSCs derived from synovial fluid and synovial tissue in both the knee and the hip joints. Our data demonstrated that there were no significant differences in colony numbers per 10 4 nucleated cells, number of nucleated cells per colony between the hip and the knee joints. In addition, the expandability of MSCs from both the knee and the hip was lost at the same time point (passage 8). However, the number of nucleated cells per milligram derived from synovium in the knee was significantly higher than that derived from that in the hip joint. Yield obtained derived from synovium in the knee joints was also not significantly, but still higher than in the hip joints. MSCs derived from synovial fluid of the hip were not capable of proliferating at passage 0. Indeed, LIVE/DEAD Viability experiments determined that both types of passage 1 cells were alive. At passage 4, cells derived from knee synovium were still alive. In contrast, reduction of viability was evident in those MSCs derived from the hip joint. These findings suggest that the potential of proliferation derived from synovial MSCs in the knee joints was superior to those in the hip joint.
Previously, it has been suggested that the potential of synovial MSCs can be affected by the presence of cartilage injury and the period between the onset of pain and date of surgery [28]. However, in this study, there was no significant difference in growth potential despite varied time frame between onset of pain and date of surgery for both the knee and the hip joints for our patients.
We also investigated the multilineage differentiation capability of MSCs derived from synovium in the hip and the knee. Our findings revealed that adipogenic differentiation potential in the knee joint was significantly higher than that in the hip joints. Interestingly, our histological findings revealed that synovial tissues from the knee joints appear to have slightly more fatty tissues than from the hip joints. These findings may be associated with why the hip synovial MSCs have lower adipogenic differentiation potential than the knee synovial MSCs. In fact, Mochizuki et al. described that the fibrous knee synovium, which is distinct from adipose synovium (infra patella fat pad), has comparable chondrogenic potential with adipose synovium and fibrous synovium has lower adipogenic potential than adipose synovium in the knee joint [29]. Our histological finding of the knee synovial tissue appears to be similar to theirs of adipose synovium.
Our findings also revealed that osteogenic differentiation potential in the knee joints was higher than that in the hip joints. The rate of von Kossa-and ALP-positive colony was not significantly different between both joints; however the expressions of RUNX2 and BGLAP in synovial MSCs derived from the knee joints at real-time PCR were significantly higher than the hip joints. On the other hand, the expressions of ALP in synovial MSCs derived from the knee joints at realtime PCR were significantly lower than the hip joints. ALP is an early marker of osteogenesis. Contrarily, RUNX2 is a late marker of osteogenesis [21,30]. Thus, our findings suggest that synovial MSCs derived from the knee differentiated more rapidly than those from the hip. This study determined that the potential of osteogenic differentiation of synovial MSCs from the knee was higher and faster than that from the hip.
Previous studies addressed the potential of the knee synovial MSCs for expansion and multiple mesenchymal lineage differentiation [11,12]. De Bari et al. first described that synovium membrane derived cells had extensive expansion potential in monolayer culture [11]. Sakaguchi et al. demonstrated that synovium from the knee was superior as a potential source of MSCs among five tissues (bone marrow, synovium, periosteum, skeletal muscle, and adipose tissue). They noted that knee synovial MSCs had the greatest expansion and differentiation ability among these tissues [12]. In addition, our previous study revealed that the shoulder synovial MSCs derived from subacromial bursa which were retained at passage 10 had the greatest differentiation potential compared to other tissues such as bone marrow, entheses, and bone [19]. In keeping with these studies, we compared potential of MSCs derived from synovial tissues in the knee and the hip joints within the same donor. Our findings indicated that synovial MSCs from the knee joints have superior capability of giving rise to adipogenic and osteogenic tissues compared to those from the hip joints. In addition, our results suggest that synovial MSCs from the knee have a higher yield compared to the hip.
In this study, we demonstrated that synovial MSCs from the knee have greater potential and differentiation capacity to those in the hip in the same donors. We showed that the nucleated cell number per milligram of sample from the knee synovium was significantly higher than that from the hip synovium. The reason why MSCs in the hip have a lower potential of MSCs than that in the knee remains unclear.
Study Limitations. There are some limitations in this study. First, our study population was small. The prevalence of simultaneous hip and knee arthroscopy is very low in the normal practice. Thus, it is extremely difficult to get the opportunity to collect synovial tissues from both joints on the same day. There is the risk of sample bias with proliferation assay due to the small sample size. Second, since this is an in vitro study, we are unable to demonstrate the potential of these MSCs in vivo. Third, we did not compare synovial MSCs to MSCs from other sources including bone marrow or adipose tissues. Fourth, histological characteristic of the knee synovium was different from the hip synovial tissue in this study. In order to eliminate the possibility of contamination of fat pad tissue or joint capsule, we obtained the tissue sample consistently by harvesting the synovium near the torn meniscus or acetabular torn labrum. Fifth, in real-time PCR study, there is possibility that interdonor variability could be larger than the difference between the knee and the hip joint.
In conclusion, our findings suggest that isolated MSCs from the hip synovial cells have self-renewal and multilineage differentiation potentials. However, adipogenesis and osteogenesis potentials of MSCs from the knees are superior to those of MSCs from the hips in the same donor. MSCs derived from the knee synovial tissue may be a better source for future mesenchymal stem cell therapy.