Auricular cartilage was obtained from young New Zealand White rabbits aged one to three months. The perichondrium was carefully removed. Samples were cut into approximately 1 mm³ pieces. Dissociation was accomplished in the culture medium, without fetal calf serum, containing 2 mg/mL collagenase (type II, Sigma) and 0.1 mg/mL testicular hyaluronidase (type IV, Sigma) for 6–8 hours at 37^°C. The resulting cell suspension was filtered through a nylon mesh, centrifuged at 300X g for 5 minutes and the cells resuspended in the culture medium. Cells were cultured as primary cultures on plastic (adherent conditions) or in suspension (aggregating conditions).

In adherent conditions, cells were cultured in Dulbecco's modified Eagle's medium (Sigma-Aldrich, St Louis, Mo) with 10% fetal calf serum (Sigma), 0.5 g/L glutamine (Sigma) and 0.5 mL/L of an antibiotic antifungal solution (Sigma) in 25 cm² tissue culture flasks at 37^°C in a water-saturated atmosphere containing 5% CO₂. The medium was changed every 2–3 days. The initial cell density was 10⁴ cells/cm². Cultured to confluence, cells were trypsinized and subcultured in a variable number of passages. Finally, they were seeded on Thermanox plastic coverslips (Nunc A/S, Roskilde, Denmark) on 24-well plates to facilitate later processing.

In aggregating conditions, chondrocyte aggregates were obtained by culturing cells in suspension on a nonadherent surface (2% agar in PBS) in conditions as above. This method prevented cell flattening and they formed small aggregates that grew for about 7 days and then retained their size and characteristics for long periods. Aggregates were made from primary and monolayer subcultured chondrocytes with similar results.

Finally, aggregates cultured for 1 to 8 weeks in suspension were either seeded on plastic or added to confluent monolayers of proliferated and dedifferentiated chondrocytes.

In addition to daily observations of cultures, cells were fixed with ethanol/acetic acid (99:1), dehydrated in ethanol and stained with hematoxylin-eosin. In other samples the extracellular matrix was stained with toluidine blue and alcian blue 8 GX (Sigma) at pH 1.0.

Indirect immunofluorescence was performed by incubating fixed (as above) and washed cultures for 2 hours with primary antibodies against type II collagen (mouse monoclonal anti collagen II, Sigma, 1 : 1500) diluted in PBS. Then samples were washed twice in PBS and incubated for 1 hours with FITC-conjugated secondary antibodies raised in goat against mouse IGM (Sigma) previously diluted 1 : 50. Washed twice in PBS, samples were mounted with a non-fluorescing mounting medium (Sigma). Control experiments were carried out in auricular cartilage and in dedifferentiated chondrocyte monolayer cultures by repeating the processing with or without primary antibody.

Auricular chondrocytes lost their characteristic round shape and grew with a fibroblast-like phenotype on plastic (Figure 1a). Dedifferentiated chondrocytes showed a typical spindle-shaped morphology with long filopodia. No special cell rearrangements occurred, nor were differentiated subpopulations present. Dedifferentiated cells showed an easy substrate attachment and a fast proliferation which formed confluent monolayers of fibroblastic cells. Neither round nor polygonal cells were observed in considerable quantity. Subsequent subcultures maintained this behavior, but a change in cell shape progressively took place. This consisted of the appearance of very large and shortly flattened cells, termed polymorphic cells (Figure 1b) owing to their irregular shape. Indirect immunofluorescence of type II collagen showed no labelling (Figure 1c).

When primary or subcultured chondrocytes were cultured in suspension, cell flattening proved impossible and many spherical aggregates of round cells appeared floating in the culture medium (Figure 1d). They grew for about 7 days, and progressively developed a dense alcian-blue-positive pericellular matrix. Aggregates were very stable in culture; when cultured for 1 to 8 weeks, no appreciable changes occurred in their appearance, except for a flattening of their peripheral cells which recalled the perichondrium in vivo (Figure 1e). The central cells were round, their pericellular matrix showed an affinity for alcian blue and toluidine blue metachromatic staining and their synthesis of type II collagen was intense (Figure 2a) like that of differentiated or redifferentiated chondrocytes. These characteristics coincide very well with those described by Stewart et al [32] in equine articular chondrocytes and Estrada et al [33] in porcine epiphyseal chondrocytes.

Aggregates of primary or subcultured chondrocytes transferred onto a plastic surface slowly adhered to it. Thereafter, peripheral cells expanded from their borders, growing radially and forming structures very similar to the chondrogenic plates of embryonic culture models (Figure 2b). Peripheral flattened cells were initially fibroblastic in shape, but they were progressively transformed into polymorphic cells, very irregular cells of greater dimensions with a high content in vacuoles. In the centers of the plates, very numerous round cells, which sometimes formed cartilaginous nodules, were surrounded by polygonal cells (Figure 2c). With regard to monolayer cultures, the cell proliferation rate apparently decreased, because cells did not cover all the available plastic surface. Many cells remained in suspension, floating in the culture medium, suggesting a decrease in substrate adhesion.

Aggregates previously kept in suspension for 1 to 8 weeks were cultured on plastic. Aggregates cultured for 1 week formed chondrogenic plates with many round and polygonal cells and few fibroblastic and polymorphic cells. Round and polygonal chondrocytes were very stable in their shape and they did not change their behavior over long culture periods of more than three months. These cells developed a refractile pericellular matrix. Round cells frequently detached from the plastic, remaining in suspension.

When the time that aggregates had been in suspension increased, the relative numbers of round and polygonal cells decreased, and fibroblastic and polymorphic cells were more abundant. Aggregates cultured for 7 and 8 weeks in suspension produced only polymorphic cells on plastic. It is interesting to note that during this time a perichondrial-like layer progressively developed and that fibroblastic and polymorphic cells, always peripheral, were able to derive from this layer.

In cultures with round and polygonal cells, type II collagen was extensively produced (Figure 2d), but in cultures with a great number of polymorphic cells the synthesis of type II collagen was negligible.

To ascertain whether the differentiation state of the aggregates has some effect on dedifferentiated chondrocytes, aggregates cultured for 1 week in suspension were added to confluent monolayers of subcultured chondrocytes. The results were similar to those described in the previous section, with the presence of many chondrogenic plates (Figures 2e, 2f). Fibroblastic cells were very scarce. Chondrogenic plates were typically made up of a central aggregate surrounded by polygonal cells and some polymorphic cells peripherally. This arrangement coincides with what was reported by Ishizeki et al [34, 35] for the culture of Meckel's cartilage chondrocytes of fetal mice. The synthesis of type II collagen was extensive and the proteoglycan staining was positive (Figure 2f).

As occurs with other chondrocyte types, the culture of auricular chondrocytes on plastic produces a dedifferentiation which turns these spherical cells into fibroblast-like cells. These dedifferentiated cells proliferate until they form a confluent monolayer. Metabolic changes in this dedifferentiated state include a changeover from type II collagen to type I collagen synthesis [14, 28, 32].

Chondrocytes have previously been cultures in suspension on agarose. Goldring [30] describes a method similar to the procedure reported here, and stated that cells first form large clumps, then these began to break up after 7–10 days, becoming a single cell suspension. No such breakup was observed in the current work; apart from different cell sources, this seems likely to be the outcome due of the agarose concentration in Goldring's method (1%) in comparison with the agar gel used here (2%).

The formation of similar aggregates in suspension cultures has been reported by Castagnola et al [5] in chick embryo tibia chondrocytes on agarose-coated dishes, by Shakibaei and De Souza [8] in mesenchymal cells from mouse limb buds cultured in alginate, by Lunstrum et al [10] in a rat mesenchymal cell line on bacteriological dishes, by Gagne et al [11] with human articular chondrocytes in an alginate suspension culture system, by Stewart et al [32] with equine articular chondrocytes, and by Estrada et al [33] with porcine epiphyseal chondrocytes. As in the present study, in all cases these aggregates showed phenotypic stability and the expression of some markers of chondrocyte differentiation. However, in most of these works, mainly with embryo chondrocytes, investigation ceased at this point.

Various methods to obtain pure chondrocyte populations using embryo cells have been described. Castagnola et al [5] report that chondrocytes from chick embryo tibiae cultured in suspension in agarose-coated dishes resumed the chondrocyte phenotype and their differentiation to hypertrophy continued; in 3–4 weeks, a nearly homogeneous population of single isolated hypertrophic chondrocytes was obtained. Shakibaei and De Souza [8] report that when isolated mesenchymal cells from mouse limb buds are cultured in alginate, only a subpopulation of differentiated chondrocytes was selected, showing a stable phenotype until the end of the culture; fibroblast-like cells became necrotic in these conditions. By dissolving alginate, cells which maintain their pericellular matrix can be cultivated in micromass to give a pure chondrocyte population. A similar method has been applied to redifferentiate adult articular chondrocytes [19].

This study describes a similar method but using auricular chondrocytes. We also report that aggregates in suspension can be transferred to plastic surfaces resulting in a considerable phenotypical stability. The results suggest that the maintenance of a differentiated phenotype on plastic occurs when chondrocytes have their pericellular matrix. We propose that cells deprived of their pericellular matrix by enzyme digestion, such as dissociated cells, attach to plastic in a dedifferentiated form. However, cells that retain their original pericellular matrix or that have reconstructed it in suspension attach to plastic with a differentiated phenotype. By other methods, Stewart et al [32] also conclude this relationship between pericellular matrix and the stable expression of the articular chondrocyte phenotype in suspension culture. Indeed, the pericellular matrix is the natural substrate of chondrocytes in vivo. Therefore, the influence of the pericellular matrix on the differentiation state of the chondrocytes hence seems to be of major importance [36].

The addition of aggregates cultured for 1 week in suspension to confluent monolayers of dedifferentiated chondrocytes produces many chondrogenic plates and cell shape became round or polygonal. The cells synthesize type II collagen and proteoglycans. Shakibaei and De Souza [8] report that mesenchymal limb bud cells cultured in alginate were selected and, while attached to their pericellular matrix, followed a chondrocyte pathway of differentiation. In aggregates of auricular chondrocytes in the present study, the pericellular matrix is visible around the cells. These results suggest that the presence of some differentiated chondrocytes surrounded by their pericellular matrix influences the differentiation state of other chondrocytes on plastic surfaces.

The maintenance of the differentiated phenotype on plastic seems to depend on the presence of some cells with their pericellular matrix. Chen et al's experiments [36] demonstrate similar behavior in chick embryo growth plate chondrocytes. Chondrocytes in organ culture, and thus with their own extracellular matrix environment, maintain their state and progress in the differentiation programme as in vivo. However, these same cells when dissociated, without extracellular matrix, and cultured in pellets, regressed in the differentiation program.

This interpretation may explain various behaviors observed in adult chondrocyte culture. Dedifferentiation on plastic would occur because cells lack their pericellular matrix, owing to the dissociation process. However, chondrocyte aggregates, with newly-formed pericellular matrix, can grow on plastic maintaining a certain differentiation. Pellet or micromass cultures obviously promote the aggregation and redifferentiation of cells in a similar manner. It is also known that dedifferentiation is attenuated by culturing chondrocytes at high density.

Patti et al [4] have reported surprising results working with adult articular chondrocytes cultured on plastic. They report that these cells in monolayer culture retained their differentiated characteristics for more than five months, including the ability to produce cartilage-specific molecules. These results seem to contradict many other studies. However, they report that as an outcome of the dissociation method used “some single or aggregated chondrocytes suspended in the digest were surrounded by remnants of territorial matrix.” Hence, it seems probable that owing to incomplete dissociation, such aggregates included in the cultures were responsible for the differentiation that these authors reported.

Consequently, a conclusion of this study is that, in addition to other known culture conditions such as cell density, chondrocyte behavior on plastic is strongly dependent on the factors outlined above. Therefore, the presence of cell aggregates should be kept in mind in chondrocyte culture studies and in tissue engineering techniques.

As an example of their potential, de Chalain et al [37] reported the production of a large amount of elastic cartilage from porcine and human chondrocytes when enzymatically isolated chondrocytes were agitated in suspension, forming chondron-like aggregates, which were further embedded in alginate and type I collagen. By implanting these constructs in nude mice, neocartilages were obtained closely resembling native auricular cartilage, both phenotypically and in matrix composition. We think that these aggregates, similar to those in the current study, might be an alternative to the use of isolated chondrocytes in both chondrocyte transplant and tissue engineering techniques. It is hoped that this report will improve the knowledge of why this success was reached.

However, further investigations focused on the regulatory mechanisms using molecular biology techniques are needed to understand these processes and their practical possibilities.