Properties and usefulness of aggregates of synovial mesenchymal stem cells as a source for cartilage regeneration

This study was approved by an institutional review board of Tokyo Medical and Dental University (No.1030), and informed consent was obtained from all subjects. Human synovium was harvested from donors during anterior cruciate ligament reconstruction surgery for ligament injury and digested in a 3 mg/ml collagenase D solution (Roche Diagnostics, Mannheim, Germany) in α-minimal essential medium (αMEM) (Invitrogen, Carlsbad, CA, USA) at 37°C. After three hours, digested cells were filtered through a 70 μm nylon filter (Becton, Dickinson and Company, Franklin Lakes, NJ, USA), and the remaining tissues were discarded. The digested cells were plated in a 150 cm² culture dish (Nalge Nunc International, Rochester, NY, USA) in complete culture medium (CCM): αMEM containing 10% fetal bovine serum (FBS; Invitrogen), 100 units/ml penicillin (Invitrogen), 100 μg/ml streptomycin (Invitrogen), and 250 ng/ml amphotericin B (Invitrogen) and incubated at 37°C with 5% humidified CO₂. The medium was changed to remove nonadherent cells one day later and cultured for 14 days as passage 0, then replated at 100 cells/cm² in a 150 cm² culture dish, cultured for 14 days and cryopreserved as passage 1. To expand the cells, a frozen vial of the cells was thawed, plated in 60 cm² culture dishes, and incubated for four days in the recovery plate. These cells were replated at 100 cells/cm² in a 150 cm² culture dish (passage 3), and cultured for an additional 14 days. These passage 3 cells were harvested and used in this study.

This study was approved by the Animal Experimentation Committee of Tokyo Medical and Dental University (No.0120296A). Wild type skeletally mature Japanese White Rabbit and GFP transgenic rabbits [11, 12] were anesthetized with an intramuscular injection of 25 mg/kg ketamine hydrochloride and with an intravenous injection of 45 mg/kg sodium pentobarbital and 150 μg/kg medetomidine hydrochloride. Synovium was harvested aseptically from knee joints of the rabbits, and digested in a 3 mg/ml collagenase type V in aMEM for three hours at 37°C. The digested cells were plated at 5 × 10⁴ cells/cm² in a 150 cm² culture dish in CCM and incubated at 37°C with 5% humidified CO₂. The medium was changed to remove nonadherent cells one day later and cultured for seven days as passage 0. The cells were then trypsinized, harvested and resuspended to be used for further assays. The cells that were transplanted in animals to be sacrificed at Day 0 and Day 14 were labeled for cell tracking by the fluorescent lipophilic tracer DiI (Molecular Probes, Eugene, OR, USA). For labeling, synovial MSCs were resuspended at 1 × 10⁶ cells/ml in αMEM without FBS and a DiI was added at a final concentration of 5 μl/ml. After incubation for 20 minutes at 37°C with 5% humidified CO₂, the cells were centrifuged at 450 g for 5 minutes and washed twice with phosphate-buffered saline (PBS) and the cells were then resuspended in CCM and cultured in hanging drops. We already reported that these cells had characteristics of MSCs [3, 6, 7, 11].

A total of 2.5 × 10⁵ synovial MSCs were trypsinized, harvested and resuspended in 35 μl of CCM, plated on an inverted culture dish lid. The lid was inverted and placed on a culture dish containing PBS. The cells were cultured at 37°C with 5% humidified CO₂ for three days in hanging drops.

Aggregates of human synovial MSCs were fixed with 2.5% glutaraldehyde in 0.1 M PBS for two hours. The aggregates were washed overnight at 4°C in the same buffer and post-fixed with 1% OsO4 buffered with 0.1 M PBS for two hours. The aggregates were dehydrated in a graded series of ethanol and embedded in Epon 812. Semi-thin (1 μm) sections for light microscopy were collected on glass slides and stained for 30 seconds with toluidine blue.

A total of 2.5 × 10⁵ human synovial MSCs cultured as a monolayer were pelleted by trypsinization and centrifugation. The pellets or aggregate of human synovial MSCs cultured for three days in hanging drops were cultured in 400 μl chondrogenic medium consisting of high-glucose Dulbecco's modified Eagle's medium (Invitrogen) supplemented with 1,000 ng/ml BMP-7 (Stryker Biotech, Boston, MA, USA), 10 ng/ml transforming growth factor-β3 (R&D Systems, Minneapolis, MN, USA), 100 nM dexamethasone (Sigma-Aldrich Corp., St. Louis, MO, USA), 50 μg/ml ascorbate-2-phosphate, 40 μg/ml proline, 100 μg/ml pyruvate, and 1:100 diluted ITS+Premix (6.25 μg/ml insulin, 6.25 μg/ml transferrin, 6.25 ng/ml selenious acid, 1.25 mg/ml bovine serum albumin, and 5.35 mg/ml linoleic acid; BD Biosciences Discovery Labware, Bedford, MA, USA). The medium was changed every 3 to 4 days for 21 days.

The pellets were embedded in paraffin, cut into 5-μm sections and stained with 1% Toluidine Blue. For immunohistochemistry, sections were treated with 0.4 mg/ml proteinase K (DAKO, Carpinteria, CA, USA) in Tris-HCl and normal horse serum after deparaffinization. Primary antibodies for type II collagen (Daiichi Fine Chemical, Toyama, Japan) and a secondary antibody of biotinylated horse anti-mouse IgG (Vector Laboratories, Burlingame, CA, USA) were employed. Immunostaining was detected with VECTASTAIN ABC reagent (Vector Laboratories) followed by 3,3'-diaminobenzidine staining.

Total RNA was extracted from human synovial MSCs in a monolayer culture, aggregates of human synovial MSCs cultured for 1, 2 and 3 days, and the pellets cultured for 7, 14 and 21 days using QIAzol (Qiagen, Hiden, Germany) and the RNeasy mini kit (Qiagen). cDNA was synthesized with oligo-dT primer from total RNA using the Transcriptor High Fidelity cDNA Synthesis kit (Roche Diagnostics) according to the manufacturer's protocol. Reverse transcription (RT) was performed by 30 minutes incubation at 55°C followed by 5 minutes incubation at 85°C. Real-time PCR was performed in a LightCycler 480 instrument (Roche Diagnostics). Primer sequences and TaqMan probes are listed in Table 1. After an initial denaturation step (95°C for 10 minutes), amplification was performed for 40 cycles (95°C for 15 seconds, 60°C for 60 seconds). Relative amounts of mRNA were calculated and standardized as previously described [13, 14].

Total RNA was extracted from human synovial MSCs in a monolayer culture, aggregates of human synovial MSCs cultured for three days. Human Genome U133 Plus 2.0 Array (GeneChip; Affymetrix, Santa Clara, CA, USA) containing the oligonucleotide probe set for more than 47,000 transcripts was used. The fluorescence intensity of each probe was quantified by using the GeneChip Analysis Suite 5.0 (Affymetrix). Gene expression data were normalized in Robust MultiChip Analysis (RMA). To analyze the data, we used hierarchical clustering using TIGR MultiExperiment Viewer (MeV) [15]. The microarray data have been deposited to the public database (GEO accession# GSE 31980).

Under anesthesia, the left knee joint was approached through a medial parapatellar incision, and the patella was dislocated laterally. Full-thickness osteochondral defects (5 mm × 5 mm wide, 1.5 mm deep) were created in the trochlear groove of the femur. A total of 5, 10, 20, 40 and 80 aggregates of autologous rabbit synovial MSCs (2.5 × 10⁵ cells/aggregate) or 25 and 100 smaller aggregates of autologous rabbit synovial MSCs (1.0 × 10⁵ cells/aggregate) suspended in PBS were transplanted to the defect. To trace the transplanted cells, DiI-labeled aggregates of autologous rabbit synovial MSCs and aggregates of allogenic synovial MSCs derived from GFP transgenic rabbit were transplanted to the defect. For the control group, the defect was left empty. All rabbits were returned to their cages after the operation and were allowed to move freely. Animals were sacrificed with an overdose of sodium pentobarbital at 1, 2, and 4 days and at 12 weeks after the operation (n = 5 at each time).

The cartilage defects were examined macroscopically for color, integrity and smoothness. Osteoarthritic changes and synovitis of the knee were also investigated. Digital images were taken using an Olympus MVX10 (Olympus, Tokyo, Japan).

The dissected distal femurs were immediately fixed in a 4% paraformaldehyde (PFA) solution. The specimens were decalcified in 4% ethylenediamine tetraacetic acid solution, dehydrated with a gradient ethanol series and embedded in paraffin blocks. Sagittal sections 5 μm thick were obtained from the center of each defect and were stained with toluidine blue and Safranin O. For fluorescent microscopic examination and terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) staining, the fixed specimens were incubated at 4°C for three hours in 5%, 10%, 15% and 20% sucrose solution, respectively. After incubation, the fixed specimens were mounted on a holder. Then 30% optimal cutting temperature (OCT) (Sakura Finetek, Tokyo, Japan) in sucrose solution was added gently into the holder. The holder was frozen in hexan chilled by dry ice and stored at -80°C. Cryosections (10 μm) were prepared with an ultracut S microtome (Reichert, Wien, Austria) and a Microm HM560 cryostat.

Histological sections of the repaired tissue were analyzed using a grading system consisting of five categories (cell morphology, morphology, matrix staining, surface regularity, cartilage thickness and integration of donor with host), which were modified from the repaired cartilage score described by Wakitani and colleagues [16], so that overly thick, regenerated cartilage could not be overestimated [6]. The scoring was performed in a blinded manner by two observers and there was no significant interobserver difference. The ratio of the safranin-O positive area over the defect was evaluated. Zeiss AxioVison software (Carl Zeiss, Oberkochen, Germany) was used for measurement of defects and safranin-O positive areas.

Aggregates of rabbit synovial MSCs were plated at 1 or 40 aggregates/well in 96-well plates (Nunc) in CCM, and incubated at 37°C with 5% humidified CO₂ for seven days without medium change. Aggregates were fixed in 4% PFA for TUNEL staining.

Human synovial MSCs were aggregated using the hanging drop technique (Figure 1A). Three days after being cultured in the drop (Figure 1B), the aggregate, consisting of 250,000 MSCs, became approximately 1 mm in diameter (Figure 1C). The aggregate was not easily broken by manipulation. Sagittal sections of the aggregates showed heart-shape as a whole (Figure 1Da). The superficial layer was composed of spindle cells parallel to the surface, whereas the deep layer was comprised of round cells both at top and bottom of the aggregate (Figure 1Db, c). Though cells positive for TUNEL staining were observed, the number was only approximately under 5% (Figure 1Dd).

To examine the sequential changes of gene expression profiles during aggregation of human synovial MSCs, microarray analyses were performed. The differences of gene profile between before and after aggregation exceeded those among donor variances (Figure 2A). The number of genes up-regulated more than five-fold was 621. The number of genes up-regulated more than 100-fold was 10, and these genes were related to hypoxia (integrin, alpha 2 (ITGA2), stanniocalcin 1 (STC1), chemokine (C-X-C motif) receptor 4 (CXCR4)), nutrient (BMP2, proprotein convertase subtilisin/kexin type 1 (PCSK1), secreted phosphoprotein 1 (SPP1), ITGA2, STC1), extracellular region (MMP1, MMP3), and cell adhesion (SPP1, ITGA2) (Table 2). The most up-regulated gene was BMP2, increased to 273 folds (Table 2). STC1 was also highly up-regulated in aggregates of synovial MSCs. The number of genes down-regulated less than one-fifth was 409, and the ontology for the genes was related to cell cycle. The microarray data are available at the public database (GEO accession# GSE 31980).

To further investigate gene expressions during aggregation of human synovial MSCs, real time RT-PCR analyses were additionally used for chondrogenesis-related genes (SRY (sex determining region Y)-box (SOX)5, -6, -9, and BMP2) and anti-inflammatory genes (TNFα inducible gene 6 (TSG-6), and STC-1) in four donors. In most cases, expressions for these genes increased sequentially (Figure 2B).

In vitro chondrogenic ability of human synovial MSCs after hanging drop culture was compared to that of MSCs after monolayer culture (Figure 3A). Aggregates of MSCs differentiated into chondrocytes as well (Figure 3B). The wet weight of pellets derived from MSCs after hanging drop culture was heavier than that of pellets derived from MSCs after monolayer culture in all four donors at 14 or 21 days (Figure 3C). Real time RT-PCR analysis showed higher expression levels of collagen (COL)2A1, aggrecan and SOX9 for pellets derived from MSCs after hanging drop culture compared to MSCs after monolayer culture at 14 and 21 days (Figure 3D). Cartilage extracellular matrix synthesis and accumulation of type II collagen were confirmed by histological analysis stained with toluidine blue and immunohistochemical analysis (Figure 3E).

At 0 days, in the case of 40 and 80 transplanted aggregates, the osteochondral defects were filled with aggregates labeled with DiI macroscopically (Figure 4A).

At four weeks, in the case of 5 and 10 transplanted aggregates, the osteochondral defect was mostly covered with a thick cartilage matrix (Figure 4B, C). In the case of 20 and 40 transplanted aggregates, the defect was partially covered with cartilage matrix. In the case of 80 transplanted aggregates, the defect was filled with only fibrous tissue, which appeared to be similar to the control (Figure 4B).

At 12 weeks, in the case of 10 transplanted aggregates, the border between cartilage and bone moved up, and thickness of the regenerated cartilage became similar to the neighboring cartilage (Figure 4B, D). In the case of 5 and 20 transplanted aggregates, the bone defect was repaired, but the cartilage defect was filled partially with cartilage matrix. In the case of 40 and 80 transplanted aggregates, the osteochondral defect was poorly repaired, similar to the control (Figure 4B). Histological score was the best and the safranin-O positive area ratio was highest in the case of 10 transplanted aggregates both at 4 and 12 weeks (Figure 4E, F).

To trace MSCs, 10 aggregates of GFP positive MSCs were transplanted into the defect. At Day 1, no GFP positive aggregates could be observed in the knee joint except the defects with a fluorescent stereomicroscope. Histologically, aggregates changed their forms but have not fused yet (Figure 5A). At four weeks, the defect was filled with cartilage matrix and the GFP positive cells were still observed both at the bottom and the center of the regenerated cartilage (Figure 5B). Regenerated cartilage consisted of both GFP positive cells and GFP negative cells.

Cell number per aggregate as well as aggregate number may be a factor affecting properties of the aggregates. To answer this question, 25 or 100 aggregates, in which an aggregate consisted of 100,000 MSCs, were transplanted into the osteochondral defect.

At four weeks, in the case of 25 transplanted aggregates, the defect was fully filled with cartilage matrix (Figure 6A), in which the result was different from the case of 20 or more aggregates, in which an aggregate consisted of 250,000 MSCs. In the case of 100 transplanted aggregates, the defect was filled with fibrous tissue, and the histological score was inferior and the safranin-O positive area ratio was smaller. (Figure 6B, C).

To clarify why transplantation of aggregates over a certain number resulted in poor outcome, viability of cells was first examined by TUNEL staining. Compared to the case of 10 transplanted aggregates, much more TUNEL positive cells could be observed in the case of 80 transplanted aggregates (Figure 6D).

Another factor might be a nutrient deprivation and in vitro analyses using aggregates of rabbit synovial MSCs were performed. Seven days after 1 or 40 aggregates were cultured in a well of 96-well plates, the medium color changed to yellow in the case of 40 aggregates, while the color remained red in the case of only 1 aggregate (Figure 6E). TUNEL positive cells were much higher in the case of 40 aggregates than in the case of only 1 aggregate.