The culture of articular chondrocytes in hydrogel constructs within a bioreactor enhances cell proliferation and matrix synthesis

Isolation of chondrocytes.

Bovine cartilage was harvested from the metacarpophalangeal joints of six 18-month-old calves. Specimens of human tissue were obtained from limbs following above-knee amputation at the Royal National Orthopaedic Hospital, Stanmore, United Kingdom. All the knees were clinically healthy, free from tumour and had not previously been exposed to radio- or chemotherapy. Human tissue was obtained from eight patients with a mean age of 23 years (17 to 29). The cartilage was obtained aseptically within six hours of collection of the tissue from the operating theatre. The study was approved by the relevant Local Ethics Committee.

The full-thickness sections of articular cartilage were collected in a culture dish containing Dulbecco’s minimal essential medium (DMEM), supplemented with 20% (v/v) fetal calf serum (FCS), 2 mM L-glutamine, 20 mM HEPES, 100 unit/ml⁻¹ of penicillin, 100 μg/ml⁻¹ of streptomycin and 50 μg/ml⁻¹ of L-ascorbic acid (DMEM + 20% FCS; all Gibco, Paisley, United Kingdom). The slices of cartilage were diced finely and incubated at 37°C for one hour in DMEM + 20% FCS + 700 unit/ml⁻¹ of pronase (BDH Merck Ltd, Poole, United Kingdom) and for 16 hours at 17°C in DMEM + 20% FCS + 100 unit/ml⁻¹ of collagenase type-XI (Sigma, Poole, United Kingdom). The supernatant containing the released chondrocytes was passed through a sieve with a pore size of 70 μm (Falcon, Oxford, United Kingdom), washed twice in DMEM + 20% FCS and finally re-suspended in DMEM + 20% FCS. The cell number and viability were determined using a haemacytometer and staining with Trypan Blue.

Formation and culture of the alginate construct.

For the preparation of alginate constructs, the cell suspension was adjusted to 20 x 10⁶ cells/ml⁻¹ for bovine and human chondrocytes. The cell suspensions were added to 4% alginate (Keltone LV; Kelvo Nutrasweet, Poole, United Kingdom) in Earle’s balanced salt solution (EBSS; Gibco) to give a final concentration of 10 x 10⁶ cells/ml⁻¹ in 2% (w/v) alginate. The cell/alginate suspension was slowly expressed through a 22-gauge needle into a solution containing 100 mM CaCI₂ in DMEM. The resultant beads were approximately spherical in shape with a diameter of approximately 3 mm.

One group of constructs was transferred to a tissue-culture dish containing 50 ml of medium and incubated at 37°C/5% CO₂ for up to 19 days. This group represented the static culture conditions with all beads lying in a single layer on a flat surface. An equal number of constructs were cultured in the 50 ml chamber of an RWV bioreactor (Synthecon, RCCS Model, Bereldange, Luxembourg). This ensured that there were equal amounts of medium per cell in both groups. A maximum of 200 constructs was used to minimise frequent collisions between the constructs during rotation of the culture chamber. The vessel was sealed and attached to the rotator base of the RWV, which was maintained at 37°C/5% CO₂ for up to 19 days. The initial speed of rotation was set to 6 to 8 rpm and, when necessary, adjusted to higher speeds to ensure that the constructs did not collide with each other or the edges of the vessel. The culture media for both groups were replaced every three days. Representative constructs were removed for subsequent analysis at selected time points.

Biochemical analysis.

Ten replicates of ten beads with bovine cells and five replicates of ten beads with human cells were taken at each time point for biochemical analysis. The chondrocyte/alginate gels were disrupted and digested by incubation in papain digest buffer, a solution containing 55 mM sodium citrate (BDH), 150 mM sodium chloride (BDH), 5 mM cysteine hydrochloride (Sigma), 5 mM EDTA (BDH) and 0.56 units/ml⁻¹ of papain (Sigma) at 60°C for 24 hours.

The total DNA was determined using the Hoechst method.^16,17 Calf thymus DNA in a solution of 55 mM sodium citrate and 150 mM sodium chloride + 0.4% (v/v) sodium dedecyl sulphate was used for preparation of the standard curve, ranging from 0 to 20 μg/ml⁻¹. Aliquots (100 μl) of standards or samples were added to appropriate wells in a fluorimetric well-plate. To each well-plate were added 100 μl of Hoechst 33258 (1 μg/ml⁻¹) in 55 mM sodium citrate and 150 mM sodium chloride and the plates were read on a fluorimeter (Fluoroskan Ascent; LabSystems, Oxford, UK) with a 355 nm excitation filter and emission at 460 nm.

Measurement of the total sulphated glycosaminoglycans (GAG) in the chondrocyte cultures was performed using the 1,9-dimethylMethylene Blue (DMB) dye method,¹⁸ which had been modified for use with an alginate system at a pH of 1.5.¹⁹ Aliquots of a papain-digested sample and a solution containing 21 μg/ml⁻¹ of DMB and 2 mg/ml⁻¹ of sodium formate and 0.5% (v/v) ethanol in distilled water (pH 1.5) were added to individual wells of a multi-well plate. The absorbance at 595 nm was determined using a microplate reader (Bio-Rad Ltd, Hemel Hempstead, United Kingdom). The concentration of GAG was determined by comparison with a chondroitin-6-sulphate standard curve (0 to 100 μg/ml⁻¹, Sigma).

The hydroxyproline content in the alginate beads was measured using an oxidation assay²⁰ which was run in a Cobas Bio centrifugal analyser (Roche Products Ltd., Welwyn Garden City, United Kingdom) at 40°C. The assay involved the oxidation of the amino acid hydroxyproline in the samples by chloramine T. The chromogens formed (pyrrole and pry-2-carboxylic acid) were coupled with p-dimethylaminobenzaldehyde to produce a chromophore, the absorbance of which could be measured at 570 nm. Standard curves were used to determine the concentration of hydroxyproline in the sample range 0 to 100 μmol/l⁻¹.

Histological examination and immunolocalisation.

Representative chondrocyte/alginate constructs collected for histological and immunohistochemical analysis were fixed in a solution containing 2% paraformaldehyde in 0.1 M NaCaCo at pH 7.4. The samples were dehydrated through a series of ethanol concentrations, cleared with xylene and embedded in paraffin wax. Sections were cut at a thickness of 8 μm using a sledge microtome and mounted on to glass slides.

For histological examination, the sections were stained using iron haematoxylin and Safranin-O to visualise cells and GAG-rich extracellular matrix. The streptavidin biotin-immunoperoxidase method was used for the immunolocalisation of collagen type-I, -II and -X, using the antibodies detailed in Table I. The sections were rehydrated and incubated in 200 μl of hydrogen peroxide in 12 ml of methanol to block the activity of endogenous peroxidase. The peroxidase-blocked sections were subsequently pre-treated in 0.1% (w/v) trypsin (Sigma) and 0.1% (w/v) CaCI₂ in distilled H₂O at 37°C for ten minutes. After pretreatment they were washed in tap water and soaked in tris-buffered saline (TBS) (0.05M TRIS, 0.15M NaCI at pH 7.6). To suppress non-specific binding of IgG, the sections were blocked with normal rabbit serum (Dako A/S, Glostrup, Denmark), diluted 1:5 in TBS for 20 minutes. The sections were then incubated with the selected primary antibody for 30 minutes at room temperature. They were washed with TBS, biotinylated rabbit anti-mouse secondary antibody (Dako A/S) was applied and they were incubated at room temperature for 30 minutes and then rinsed. Streptavidinbiotin complex (Dako A/S) was applied for 30 minutes, followed by the liquid substrate solution containing 0.7 mg/ml⁻¹ of 3′ 3′-diaminobenzidine tetrahydrochloride, 0.7 mg/ml⁻¹ of urea hydrogen peroxide and Tris buffer 0.006 M (Sigma) was prepared. The sections were counter-stained with Harris’s haematoxylin, washed in tap water and differentiated in acid alcohol before mounting and visualisation.

Statistical analysis.

The comparison between the biochemical data obtained from static and RWV cultures was performed using the unpaired Student’s t-test. A p value of < 0.05 represented a statistically significant difference.

Biochemical analysis.

The DNA content of bovine chondrocyte-seeded alginate constructs cultured for up to 19 days in static conditions and within the RWV is shown in Figure 1a. The initial DNA concentration was 48.2 μg/ml⁻¹ of alginate. DNA values for the bovine chondrocyte-seeded constructs increased slowly during the initial stages of culture, reaching a level by day seven of 73.0 μg/ml⁻¹ and 92.2 μg/ml⁻¹, in the static and RWV cultures, respectively. Thereafter, the DNA values increased rapidly, with the increase being greatest in the RWV specimens. Accordingly, on day 14, the concentrations were 144 μg/ml⁻¹ for static cultures and 201 μg/ml⁻¹ for RWV cultures, the differences being statistically significant (p = 0.01). The concentrations continued to increase in the RWV culture up to day 17, but decreased in the static culture (Fig. 1a).

Fig. 1a, Fig. 1b

Graphs showing the DNA content of a) bovine chondrocyte-seeded alginate beads and b) human chondrocyte-seeded alginate beads, cultured for 19 days in a dynamic rotating wall vessel (RWV) bioreactor and a static petri culture dish.

Corresponding DNA values within the alginate constructs which had been seeded with human chondrocytes are shown in Figure 1b. The initial concentration of DNA was 47.6 μg/ml⁻¹ of alginate. There was little change in the DNA concentration in the RWV cultured constructs up to day 12. The corresponding static cultures showed a decreased DNA concentration after seven days in culture compared with the initial values. Subsequently, the concentrations increased, so that by day 19 the values were 86.8 μg/ml⁻¹ and 91.2 μg/ml⁻¹ for the static and RWV cultures, respectively. The differences between the two cultures were not statistically significant (p = 0.01).

GAG values within alginate constructs seeded with bovine chondrocytes and cultured for up to 19 days are given in Figure 2a. The initial values measured within the constructs were 127 μg/ml⁻¹ of alginate. They increased slowly during the first week of culture, but there was no significant difference between the static condition (307 μg/ml⁻¹) and the RWV (285 μg/ml⁻¹) at day seven. Thereafter, the rate of increase was greatly enhanced in the RWV cultured constructs, compared with the static model. By day 19, the values were 1.23 mg/ml⁻¹ and 2.33 mg/ml⁻¹ for the static and RWV conditions, respectively. The difference in the GAG levels between the two systems was highly significant at day ten and thereafter (p = 0.06).

Fig. 2a, Fig. 2b

Graphs showing the total GAG content of a) bovine chondrocyte-seeded alginate bead constructs and b) human chondrocyte-seeded alginate bead constructs, cultured for 19 days in a rotating wall vessel (RWV) bioreactor and a static petri culture dish.

The corresponding GAG values for alginate constructs seeded with human chondrocytes are illustrated in Figure 2b. The concentration in the constructs at the start of the experiment was 274 μg/ml⁻¹ for both groups and increased steadily throughout the culture period for both culture conditions, although the rate of increase was significantly greater in the RWV cultured constructs. By day 19, the GAG values were 996 μg/ml⁻¹ and 1664 μg/ml⁻¹ for static and RWV cultures, respectively (p = 0.005).

The content of hydroxyproline within the constructs seeded with bovine chondrocytes are given in Figure 3a. The initial hydroxyproline measured was 80 μg/ml⁻¹. The concentration gradually increased during the week in culture. There was a rapid increase in the rate of production in both culture conditions from day seven. At day 14, the content measured 800 μg/ml⁻¹ and 1100 μg/ml⁻¹ for the static and RWV conditions, respectively. The difference in concentration between the two culture conditions was significant at day 14 and thereafter (p = 0.01).

Fig. 3a, Fig. 3b

Graphs showing the total hydroxyproline content of a) bovine chondrocyte-seeded alginate bead constructs and b) human chondrocyte-seeded alginate bead constructs, cultured for 19 days in a rotating wall vessel (RWV) bioreactor and a static petri culture dish.

The corresponding values for human chondrocytes are illustrated in Figure 3b. The hydroxyproline concentration was 70 μg/ml⁻¹ for both groups at day zero and increased steadily in both culture conditions, although the rate of increase was significantly greater in the RWV cultured from day seven onwards. By day 19, the values were 800 μg/ml⁻¹ and 1122 μg/ml⁻¹ for static and RWV cultures, respectively (p = 0.01).

Histological examination and immunolocalisation.

Figure 4 shows the histological findings in the central (Figs 4a and b) and peripheral (Figs 4c and d) regions of bovine chondrocyte-seeded alginate constructs, maintained in static culture (Figs 4a and c) or within the RWV (Figs 4b and d), for 14 days. In the peripheral region of constructs maintained in static culture, the cells were well dispersed, in groups of up to four and had a rounded morphology. A pericellular region of GAG-rich extracellular matrix, approximately five to 10 μm in thickness surrounded most cell groups, as indicated by staining with Safranin O (Fig. 4c). At the extreme periphery, however, a layer of flattened cells was present which formed a capsule around the construct. By contrast, at the centre of the construct no clumps of cells were observed. Safranin-O staining was weak and present only in the immediate pericellular space (Fig. 4a). When cultured in the RWV, an increase in the intensity and extent of Safranin-O staining was noted in all regions when compared with static conditions (Figs 4b and d). The three regions of matrix (pericellular, interterritorial and territorial) were identifiable and in some locations, neighbouring clumps of interterritorial matrix had combined. Moreover, cell clusters often consisted of ten or more cells in the peripheral region (Fig. 4d) and up to four cells in the central region (Fig. 4b). The cells at the periphery were larger and had enhanced Safranin-O staining compared with those in the central region. No capsule of flattened cells was present surrounding the constructs cultured in the RWV bioreactor.

Fig. 4a, Fig. 4b, Fig. 4c, Fig. 4d

Photomicrographs of bovine chondrocytes in alginate beads at day 14 in Safranin-O stain (x 20 magnification). a) in static culture, centre of the bead; b) in rotating wall vessel (RWV) bioreactor, centre of the bead; c) in static culture, peripheral edge of the bead; d) in RWV bioreactor, peripheral edge of the bead.

The corresponding histological images for human chondrocytes in alginate constructs are shown in Figure 5. In the static cultures, diffuse Safranin-O staining was observed within the peripheral region with a slight increase in staining intensity around the chondrocytes (Fig. 5c). These showed a large variation in size and were distributed singly or in pairs. In some cases, flattened cells were present at the extreme periphery of the construct. Figure 5a shows that human cells at the centre were associated with only a small amount of extracellular matrix and had minimal cell division. Culture within the RWV bioreactor enhanced Safranin-O staining in both the peripheral and central regions, compared with static culture (Figs 5b and 5d) and the staining was more cell-associated. An increase in the size of cell clusters was also noted, with groups of up to four cells present in both the peripheral and central regions. All the cells had a rounded morphology and the capsule of flattened cells was absent.

Fig. 5a, Fig. 5b, Fig. 5c, Fig. 5d

Photomicrographs of human chondrocytes in alginate beads at day 14 in Safranin-O stain (x 20 magnification). a) in static culture, centre of the bead; b) in rotating wall vessel (RWV) bioreactor, centre of the bead; c) in static culture, peripheral edge of the bead and d) in RWV bioreactor, peripheral edge of the bead.

Figure 6 shows collagen type-II immunolocalisation within constructs seeded with bovine (Figs 6a and b) and human (Figs 6c and d) chondrocytes cultured in static conditions (Figs 6a and c) or within the RWV bioreactor (Figs 6b and d). Bovine chondrocytes cultured in static conditions had limited type-II staining, primarily associated with rounded cells at the periphery of the construct (Fig. 6a). The cells at the surface of the construct with a flattened morphology did not express collagen type-II (Fig. 6a). In the RWV culture the staining for collagen type-II was of greater intensity and distributed more extensively throughout the construct when compared with static conditions. In static cultures, human chondrocytes stained poorly for collagen type-II (Fig. 6c), while staining was more intense and widely distributed within the cultures in the RWV bio-reactor (Fig. 6d).

Fig. 6a, Fig. 6b, Fig. 6c, Fig. 6d

Photomicrographs showing collagen type-II in a) bovine static culture; b) bovine rotating wall vessel (RWV); c) human static culture and d) human RWV.

Type-I collagen was expressed only by cells with a flattened morphology, which formed a capsule surrounding both bovine and human chondrocyte-seeded constructs cultured in static conditions (Fig. 7). No type-I collagen expression was observed for constructs maintained in the RWV bioreactor (data not shown since staining did not appear). Type-X collagen staining was not detected in any constructs, although the positive control (growth plate cartilage) had intense staining in the calcified cartilage region (data not shown).

Fig. 7a, Fig. 7b

Photomicrographs showing collagen type-I staining in a) bovine static culture and b) human static culture.