Chondrocytes from OA patients were obtained from macroscopically intact cartilage areas of 29 patients undergoing knee joint implant surgery. Samples were taken from the intercondylar femoral notch (fossa intercondylica). The average age of the OA patients at the time of surgery was 67.2 ± 10.1 years (minimum 46 years, maximum 89 years). To compare the status of these cells with cells that are actually used for ACT, chondrocytes obtained from human articular cartilage biopsies from the femoral notch of 41 patients undergoing ACT were included in this study. All procedures followed the guidelines for ACT to treat chondral defects 6. ACT surgery was performed as described previously 3. The average age of the patients at the time of ACT was 32.3 ± 10.0 years (minimum 16 years, maximum 57 years).

Since all the chondrocytes from the ACT patients had to be used for expansion and transplantation, no ACT cells were available for analysis ex vivo. As a surrogate for ACT ex vivo controls, chondrocytes were obtained from the cartilage of six knee joints of individuals without any osteoarthritic symptoms (36.6 ± 12.5 years, minimum 23 years, maximum 50 years) either post mortem (n = 1) or after amputation (n = 5). The study was approved by the local ethics committee.

Cartilage samples, excluding the mineralized cartilage and the subchondral bone, were washed twice in PBS (BioWhittaker, Verviers, Belgium) and then minced. Extracellular matrix was enzymatically degraded overnight by incubation in DMEM/Ham's F12 medium (BioWhittaker; Verviers, Belgium) containing 2.5 mg/ml type II collagenase (Roche, Mannheim, Germany) and 5% serum at 37°C. Cell culture medium for ACT chondrocytes was supplemented with autologous serum, culture medium for OA chondrocytes with human AB serum. Isolated chondrocytes were resuspended by pipetting up and down several times and then filtered through a 100 μm mesh to remove undigested cartilage fragments and extracellular matrix debris. After centrifugation, the cells were resuspended in DMEM/Ham's F12 cell culture medium supplemented with either 10% autologous or AB serum and plated in cell culture flasks (BD Falcon, Heidelberg, Germany) at an initial density of 1,500 cells/cm². At this point, some of the cells were harvested to provide ex vivo cells.

Chondrocytes were cultured at 37°C in humidified atmosphere containing 5% CO₂. The cells were harvested after 10 to 12 days of expansion by trypsin-EDTA (BioWhittaker) treatment. Cell yields and viability were monitored by trypan blue staining using a Neubauer hematocytometer. At this time, cells were removed to determine gene expression patterns after primary expansion (P0), used for in vivo experiments, or cultured for an additional 12 to 14 days to provide first subculture (P1) cells. All procedures were performed according to the good manufacturing practice guidelines required for tissue engineering.

RNA was extracted and isolated from chondrocytes using the RNeasy mini kit according to the manufacturer's instructions (Qiagen Inc., Valencia, CA, USA). To isolate RNA from the cell-seeded scaffolds that were implanted into SCID mice, the scaffolds were frozen in 350 μl RLT buffer (Qiagen RNeasy Mini kit) supplemented with 10 μl/ml β-mercaptoethanol. Scaffolds were then homogenized using a micropestle (Eppendorf, Hamburg, Germany) and samples were frozen at -80°C until further isolation.

Complementary DNA (cDNA) was obtained by reverse transcription of 1 μg total RNA using oligo-dT primers and MuMLV reverse transcriptase (BD Clontech, Heidelberg, Germany). Reverse transcription was performed in a total volume of 20 μl at 42°C for 1 h in a thermocycler (Whatman Biometra, Göttingen, Germany). Expression of mRNA/cDNA levels was determined by quantitative real-time RT-PCR (qRT-PCR; LightCycler^®, Roche) using specific target primers (Table 1) and FastStart DNA SybrGreen reagents (Roche) according to the protocols provided. The amplification of cDNA was performed in 35 PCR cycles: after 5 initial cycles (95°C 10 s, 68°C 10 s, 72°C 16 s, temperature transition rate 20°C/s) the annealing temperature was dropped in consecutive cycles to 60°C with a step size of 0.5°C.

To monitor the specificity of the amplification, melting curve analysis was performed after each PCR. In addition, some samples were separated by electrophoresis and visualized on agarose gels to confirm the size and purity of the PCR products. Amplification of glyceraldehyde-3-phosphate dehydrogenase (GAPDH) encoding cDNA and serial dilutions of a recombinant standard with a known DNA concentration (Roche) served as controls in each PCR. Data show the mean of the mRNA expression levels of the gene investigated normalized by the respective GAPDH signal and recombinant standard in each individual sample and PCR reaction. To show the relative copy numbers of the different genes investigated (ranging from more than 10⁶ to less than 1 copy/μl cDNA) qRT-PCR data are shown on a log scale. This required an adjustment of all normalized values by a factor of 100,000. Statistical evaluation of the data was performed by a Mann-Whitney U test. Groups were considered statistically different when the probability values p were equal to or smaller than 0.05.

To investigate the capability of OA chondrocytes to form cartilage under in vivo conditions, primary culture cells (P0 cells) from three osteoarthritic donors (OA donor 1, age 78 years; OA donor 2, age 68 years; OA donor 3, 50 years) and two healthy donors (healthy donor 1, age 50 years; healthy donor 2, age 42 years) were harvested by mild enzymatic detachment, washed, counted, resuspended in cell culture medium and seeded onto a biphasic collagen matrix (Jotec AG, Hechingen, Germany). The scaffold consisted of a bovine collagen membrane on one side and a porous collagen sponge on the other side. The sponge side of the scaffold was seeded with 1 × 10⁶ or 3 × 10⁶ OA chondrocytes/cm² or 1 × 10⁶ healthy chondrocytes/cm².

The cell-scaffold constructs were then cultured in vitro for an additional 4 days, after which they were implanted subcutaneously into female CB-17/Lcr SCID mice aged 10 to 12 weeks (Charles-River Wiga, Sulzfeld, Germany; n = 3 per group). The mice were anesthetized using ketamine and xylazine (1 ml 10% ketamine (WDT eG, Garbsen, Germany) and 1 ml xylazine (Rompun^®, WDT eG) in sterile PBS; 0.1 ml/10 g body weight subcutaneously). Two scaffolds were implanted subcutaneously at the back of each mouse through a small incision in the neck region. Empty scaffolds were used as controls. The mice were kept in isolator cages in an air-conditioned specific pathogen free facility on an unrestricted diet. After 8 weeks the mice were sacrificed using CO₂, and the constructs were harvested and fixed in 10% formalin buffered with 0.1 M phosphate buffer (pH 7.4). All procedures were approved by the local animal care committee.

In an additional experiment, scaffolds seeded with cells from four OA patients or three ACT patients were implanted into the mice and harvested after eight weeks for mRNA isolation.

After fixation, the constructs were embedded in Tissue Tec compound (Sakura, Zoeterwoude, The Netherlands) and 7 μm sections were cut with a cryomicrotome (Jung/Leica Instruments, Nussloch, Germany). To determine if synthesis of cartilage-like tissue had occurred, we stained sections with safranin O/fast green to show the presence of proteoglycans. Type I and type II collagen was also visualized using standard immunohistochemistry. Type I collagen was detected using the 1-855 monoclonal antibody (IgG2a, ICN Pharmaceuticals, Aurora, OH, USA), type II collagen using the II-II6B3 monoclonal antibody (IgG1, kappa light chain) 12 followed by a biotin-labeled horse anti-mouse serum (Vector, Burlingame, CA, USA). A biotin-streptavidin detection system (Vectastain Elite Kit, Vector) was used according to the manufacturer's recommendations. The II-II6B3 antibody was obtained from the Developmental Studies Hybridoma Bank maintained by the Department of Pharmacology and Molecular Sciences, Johns Hopkins University School of Medicine, Baltimore, MD, and the Department of Biological Sciences, University of Iowa, Iowa City, IA, under contract NO1-HD-2-3144 from the NICHD.

Hypertrophic chondrocytes were detected by staining for alkaline phosphatase (ALP). Sections were washed in ALP buffer (0.1 M Tris-HCl, 0.1 M NaCl, 5 mM MgCl₂, pH 9.0) and then incubated with 3.5 μl 5-bromo-4-chloro-3-indolyl phosphate (50 mg/ml, Sigma Aldrich, Taufkirchen, Germany) and 4.5 μl nitroblue tetrazolium (50 mg/ml, Sigma-Aldrich) per ml ALP buffer. The reaction was stopped by washing with PBS.

In the first set of measurements, gene expression patterns were investigated in cells directly after isolation from cartilage. As all of the chondrocytes obtained from the ACT patients were needed for expansion and subsequent transplantation, ex vivo ACT chondrocytes were not available for experimental investigation. Instead, cells from healthy cartilage served as surrogates. Ex vivo chondrocytes from healthy donors showed a prominent type II collagen signal by qRT-PCR and a somewhat weaker mRNA expression than OA chondrocytes (Figure 1). Type I collagen encoding mRNA was expressed to a lesser extent, resulting in very low type I to type II collagen ratios in both groups. The expression of aggrecan mRNA was high in both chondrocyte populations ex vivo. In ex vivo OA chondrocytes it exceeded the mRNA amounts found in chondrocytes from healthy donors (Figure 1), indicating that cells from both groups were in a highly differentiated state. This was confirmed by the low expression of ALK-1, a marker for chondrocyte dedifferentiation, in both groups (Figure 1). However, despite having similar collagen and aggrecan expression patterns, significant differences in IL-1β mRNA levels could be seen between healthy and OA cells: in OA chondrocytes, ex vivo IL-1β levels were more than 8,000 times (p < 0.05) higher than in the healthy controls (Figure 1).

Interestingly, the high IL-1β expression observed ex vivo in OA chondrocytes dropped strongly (1,448-fold) after 10 to 12 days of in vitro culture to expression levels only slightly higher than the IL-1β expression of healthy chondrocytes ex vivo or of ACT chondrocytes after primary culture. Although the increase in type I collagen and the decrease in aggrecan expression levels during culture suggested a slight dedifferentiation of the OA chondrocytes, culture of the OA chondrocytes also resulted in a four-fold increase in type II collagen expression compared to the ex vivo values (Figure 2). These levels were almost nine times higher than those in ACT cells cultured using the same protocol. Since the type I collagen expression levels differed only slightly between OA and ACT chondrocytes (Figure 2), this led to a significantly better type I/type II collagen ratio in OA chondrocytes than in the OA cells. Combined with the slightly lower ALK-1 expression in the OA cells, this suggests that the phenotypic state of OA chondrocytes is at least comparable to, if not better than, that of ACT cells at the stage where the latter are used for transplantation back into the patient.

Since defect sizes in OA cartilage are expected to be larger than those currently treated using the ACT method, the number of cells required for ATC will be greater as well. Therefore, we further expanded the OA chondrocytes and analyzed the gene expression patterns in first subculture cells (P1). We found that the expression of type II collagen mRNA was much weaker (23-fold; Figure 3) in P1 OA chondrocytes than in P0 cells. Although the expression of type I collagen remained constant, this still led to a 270-fold increase of the type I to type II ratio in P1 OA cells compared to the P0 OA cells, an indicator for dedifferentiation of the cells. It is interesting to note that in P1 OA chondrocytes, IL-1β expression decreased further to levels found in healthy chondrocytes ex vivo and in primary culture ACT chondrocytes (Figure 3).

To investigate whether P0 OA chondrocytes might be employed for the regeneration of cartilage defects by ACT, we seeded such cells onto biphasic collagen scaffolds and implanted them subcutaneously into immune deficient mice. This scaffold consists of a very slowly degrading membrane and a porous region (sponge) that is normally replaced by cartilage-like tissue within eight weeks. After eight weeks of implantation, no cartilage formation could be observed in the empty control scaffolds (Figure 4a,b). Some scattered cells of murine origin were present inside the scaffold (data not shown). In scaffolds seeded with the lowest tested density of chondrocytes (1 × 10⁶ cells/cm²), cartilage-like tissue containing proteoglycans (Figure 4c) and type II collagen (Figure 4d) was formed only in scaffolds seeded with cells from OA donor 1. In scaffolds seeded with cells from the other two OA donors (Figure 4g,i) only isolated cells staining positive for type II collagen could be observed, and there was insufficient cartilage formation. However, when chondrocytes were seeded at a higher density (3 × 10⁶ cells/cm²), cartilage was generated by cells from all three donors (Figure 4e,f,h,j), in levels similar to those found in the scaffolds seeded with 1 × 10⁶ healthy chondrocytes/cm² (Figure 4k; one sample of three healthy donors). In these samples the spongy part of the scaffold was completely replaced by hyaline-like cartilage, as indicated by the presence of cells with a round, chondrocyte-like morphology embedded in proteoglycan – (Figure 4e) and type II collagen-rich tissue (Figure 4f,h,j). Hardly any type I collagen could be detected in the newly formed cartilage (Figure 4l). The absence of cell clustering (insert in Figure 4e), ALP activity (Figure 4m) and of large, hypertrophic chondrocytes (insert in Figure 4e) suggest that these cells show no inclination to become hypertrophic or to retain OA characteristics.

To investigate the regenerative potential in OA chondrocytes in comparison to ACT cells in more detail, cells from both cohorts were expanded in primary culture, seeded onto collagen scaffolds, and implanted subcutaneously into SCID mice as described above. After eight weeks in situ, the scaffolds were harvested to determine the gene expression patterns by qRT-PCR. Samples derived from ACT patients showed no significant differences in mRNA expression levels of any of the genes investigated compared to cells from OA patients (Figure 5). Interestingly, the expression of IL-1β mRNA remained below qRT-PCR detection levels in all the samples. Furthermore, only low levels of type X collagen mRNA expression could be detected, suggesting that the implanted cells did not become hypertrophic. We conclude from these data that chondrocytes harvested from intact sites of articular cartilage of OA patients are not significantly different with respect to the factors investigated in this study and seem to retain at least some regenerative potential.