Patient characteristics
This study was approved by the institutional ethical review board of our Hospital (Rapid review of scientific research projects for use of discarded biological material), and informed consent was obtained from all donors. Twenty-eight patients (9 male and 19 female, mean age, 63.6 years [range, 53-73 years]; mean body mass index, 26.0kg/m2 [range, 22.7-30.8 kg/m2]; mean disease duration 7.3 years [range, 3-15 years]) (Supplementary Table 1) who were diagnosed with late-stage idiopathic KOA according to the criteria of American College of Rheumatology [24] with varus malalignment of the lower extremity and scheduled for elective TKA were recruited. Radiographs exhibited a relatively spared lateral femoral compartment (joint space>3mm). Cartilage morphology was scored according to the Whole-organ magnetic resonance imaging score [25] (mean cartilage scores 19.4 [range, 12.0-25.0] for medial femorotibial joint; mean cartilage scores 6.2 [range, 3.0-10.0] for lateral femorotibial joint) (Supplementary Table 1). Patients were excluded if they had secondary arthritis related to systemic inflammatory arthritis or if their history included previous systemic or intra-articular injection glucocorticoids, prior ipsilateral knee surgery, knee injury, infection, or osteonecrosis. There are 28 patients including in our study. 18 patient specimens were used for histopathology experiments and 10 other patient specimens were used for cell isolation. Among 18 patient specimens, 6 patient specimens were too hard to completely decalcified, thus 12 were used for HE staining, among them, 8 randomly selected patient specimens were used for CD105/CD271 staining.
Isolation, expansion and identification of mCPCs
During the arthroplasty procedure, an osteochondral specimen of the tibial plateau was harvested with the initial proximal tibial cut. Samples of Outerbridge grade 1-2 cartilage were obtained from the lateral tibial plateau, and samples of grade 3-4 cartilage were obtained from the medial tibial plateau (n=10 donors). Grade 1-2 cartilage includes cartilages with an intact surface (grade 1) and minimal fibrillation (grade 2), and grade 3-4 cartilage includes cartilage with fissures to subchondral bone [26]. The methods used to harvest the CPCs have been described in previous studies [5, 12, 27, 28] with minor modifications. In brief, the cartilaginous tissues were separated from the osteoarthritic articular cartilages without contaminated subchondral bones and were minced into pieces (about 1.0mm×1.0mm×1.0mm, Fig. 1B), and then digested in 0.1% collagenase II (Sigma) for 2 hours. The released cells were abandoned and the digested cartilage chips were incubated in alpha-minimal essential medium (α-MEM) with 10% vol/vol fetal bovine serum (FBS) (Invitrogen Life Technologies) at 37℃ in an atmosphere of 5% CO2. The mCPCs outgrew from cartilage chips within 10 days and the adhesive cells rapidly reach 60-80% confluence in another 5 days. Importantly, the cartilage chips were retained and maintained until passage 3 to mimic the stem/progenitor niche ex vivo and allow more CPCs outgrowth. The morphological characteristics of CPCs were observed with reverted light microscope (Olympus BX71).
Flow cytometry analysis
The cell surface antigen profile of paired mCPCs (n = 6 donors) was analyzed by flow cytometry. mCPCs at passage 3 were harvested by trypsin digestion, and antibodies were stained individually (phycoerythrin (PE)-conjugated monoclonal antibodies against human CD29, CD44, CD73, CD166, fluorescein isothiocyanate (FITC)-conjugated monoclonal antibodies against human CD45, CD90, CD271, and allophycocyanin (APC)-conjugated antibodies against CD31, CD105, eBio-Science) for 30 min in the dark at 4℃. After 2 washes with phosphate-buffered saline (PBS). Events were collected by flow cytometry with a FACScalibur system (BectoneDickinson), and the data were analyzed using FlowJo 7.6 software.
Growth kinetics and CCK-8 assay
The growth kinetics were determined using the trypan blue exclusion cell count method for hemocytometer cell counting [29]. Briefly, paired mCPCs (n = 6 donors) were cultured in 48-well plates at 2×104 cells/well and harvested every 3 days for hemocytometer cell counting during a period of 19 days. The Cell Counting Kit 8 (CCK-8, Dojindo, Japan) assay was conducted according to a previous study [30]. In brief, CPCs at passage 4 were seeded in 96-well plates (1×103/well, 5 wells in each group) and maintained in culture medium, and the CCK-8 solution was added at a ratio of 100µl/ml and incubated at 37°C for 1 hour. The absorbance was measured at a wavelength of 450 nm on days 1, 4, 7, 10, 13, 16 and 19.
Colony-forming unit fibroblast formation (CFU-F) assay
Passage 4 paired mCPCs (n = 6 donors) in each group were adjusted to different cell numbers (1×103, and 5 ×103cells/well). Aliquots of cell suspensions were added to 6-well culture plates and were maintained in culture for 10 days. Crystal violet was used to stain the colonies, and their vertical gross appearances were imaged by digital photography.
Evaluation of the multipotency of mCPCs
Osteogenic, adipogenic, and chondrogenic differentiation was assayed. The previously reported protocols for CPCs differentiation were used with minor revision in the current study [23]. Briefly, for osteogenic differentiation, cells were harvested and incubated in osteogenic induction medium (10 mM of glycerol-2-phosphate, 0.1 µM of dexamethasone and 20 µM of ascorbic acid) for 14 or 28 days. The osteogenic activity was assessed at day 14 by alkaline phosphatase (ALP) staining and at day 28 by Von Kossa staining, respectively. For adipogenic differentiation, CPCs were cultivated at 1×104 cells/well in 48-well cell culture plates, adipogenic induction medium (1 µM of isobutylmethylxanthine and 10−3µM of dexamethasone) was supplemented, and Oil red O staining (day 14) was performed to assess the adipogenic potency. For chondrogenic differentiation, 4×105 CPCs were centrifuged in polypropylene tubes to form a pelleted micromass and maintained in chondrogenic induction medium consisting of α-MEM supplemented with 107M of dexamethasone, 1% (vol/vol) insulin-transferrin-sodium selenite, 50μM of ascorbate-2 phosphate, 1mM of sodium pyruvate, 50μg/ml (wt/vol) of proline, and 20ng/ml (wt/vol) of transforming growth factor (TGF-β3). On day 28, the pellets were fixed and sectioned. The development of chondrocytes and accumulation of the cartilage matrix were evaluated by hematoxylin-eosin, toluidine blue and Safran O staining. The expression of Sox-9 (SRYtype high-mobility group box-9) and Col-II (collagen type II) were detected by immunohistochemical assays. The images were captured using a microscope under brightfield mode. Chondrogenesis was also analyzed according to a previously published histological pellet scoring system[31]. Data were obtained from six paired samples, with each repeated in triplicates.
Histologic and immunohistochemical analysis
The osteochondral specimens of initial proximal tibial cut during the arthroplasty procedure were also collected for histologic immunohistochemical analysis (n=18 donors). Samples were placed in 10% formalin before processing. For each patient, separate lateral and medial tibial plateau pieces were decalcified using 10% ethylenediaminetetraacetic acid (EDTA, Sigma) for 3-4 months and then mounted on paraffin blocks. Decalcified tissue specimens were stained with hematoxylin and eosin. Immunohistochemistry for CD105 and CD271 (NGF receptor) staining was performed. Mouse anti-human CD105 and CD271 monoclonal antibody (Abcam) was used at a dilution of 1:50. Digital image analysis was performed to evaluate relative cartilage damage (including the cartilage–bone interface) and CD105+ and CD271+ cells in vivo distribution. For each sample, the whole tissue area was scanned using an OlympusX71 microscope under brightfield mode depending on the size of the section, 2-5 images were captured for the cartilage area (including the cartilage–bone interface).
OA synovial fluid-mediated migration of mCPCs
Migration of paired mCPCs (n = 5 donors) on stimulation with OA pro-inflammatory synovial fluid [32] (from three symptomatic idiopathic KOA patient) was analyzed in 24-well transwell plates (8μm pore size of polycarbonate membranes, Corning) as described previously[33]. In brief, 5×104 mCPCs in serum-free α-MEM medium were seeded in the upper wells. The lower wells were filled with 0%, 20% and 40% OA synovial fluid. After incubated at 37°C for 10 or 20 hours, cells that migrated through the polycarbonate membrane were fixed with acetone/methanol (1:1, vol/vol). Non-migrating cells on top of the membrane were removed. Migrated cells were stained by 4′, 6-diamidino-2-phenylindole (DAPI) and crystal violet and counted microscopically. Three representative photographs (left, right, and central) of each well were taken, migrated cells per picture were counted using Image J (National Institutes of Health, Bethesda, MD), and the total number of migrated cells was extrapolated to the total well and the migration rates were calculated.
Real-time quantitative polymerase chain reaction (RT-qPCR)
RT-qPCR was performed to further evaluate their multilineage differentiation and RNA Sequencing validation. After maintaining in osteogenic, adipogenic, and chondrogenic differentiation media at a density of 5×104 cells/well in 6-well cell culture plates for 10 days, the total RNA was extracted using Trizol reagent (Fermentas), and reverse transcribed using an mRNA Selective PCR Kit (TaKaRa) according to the manufacturer’s instructions. Human runt related transcription factor 2 (RUNX2), osteocalcin (OCN), CCAAT/enhancer binding protein alpha (CEBP/α), peroxisome proliferator-activated receptor gamma (PPARγ), sex determining region Y-box 9 (Sox-9), and collagen type II (Col-II) cDNA were amplified by real-time PCR using a SYBR PCR Master Mix Kit (Sigma-Aldrich). The primers were synthesized by Invitrogen (Shanghai, China), and the sequences are shown in Supplementary Table 2. The mRNA levels were normalized to the value of β-actin or RPL13a (housekeeping genes for Sox-9 and Col-II only) [34]. Mean fold changes were calculated. Data presented are mean of six different donors, with each repeated in triplicates.
mRNA expression profile of mCPCs by RNA sequencing analysis
We used equal amounts of total RNA from each of 6 patients’ paired mCPCs from grade 1-2 and grade 3-4 osteoarthritic cartilage. The RNA sequencing was performed by GeneWIZ Technology (Suzhou, China). Briefly, the quality control of Gene expression profile analysis was performed by using Agilen bioanalyzer 2100 system. The mRNAs were captured by NEBNext Poly (A) mRNA magnetic isolation module. Library construction was conducted by NEBNext ultra RNA libray RNA PREP kit for Illumina. Library purification were conducted by Beckman Agencourt AMPure XP beads. Library quantification and results verification were performed by Agilen bioanalyzer 2100 &Qubit system. CBOT clustering and Hiseq were respectively performed by using TruSeq PE cluster Kit V4 and TruSeq SBS Kit V4-HS. Bioinformatics analysis was performed according to the manufacturer’s protocols. We then selected relative expression of genes associated with OA pathogenesis (involved in mesenchymal stem cell [MSC] tripotentiality, collagen metabolism, chemotaxis, angiogenesis, and control of osteoclast activation and other genes). We clustered the significantly increased and decreased genes according to various biological processes, cellular component, molecular function and analyzed the differentially expressed genes in the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways [35]. Expression of arbitrarily selected dysregulated genes were validated by RT-qPCR (n= 6 donors).
Statistical analysis
Data were presented as mean values and standard deviation (SD). The normal distribution of data was confirmed with the Kolmogorov-Smirnov test. As for normally distributed data (flow cytometric measurements, CFU-F assays, growth kinetic parameter, gene expression, and migration rates between mCPCs from paired grade 1-2 and grade 3-4 cartilage), a paired t-test was employed, for ordinal grading data such as the pellet histological scores, a Wilcoxon signed-rank test was applied. A P-value <0.05 was considered statistically significant. All tests were performed using IBM SPSS Statistics 20.0