Time evolution of in vivo articular cartilage repair induced by bone marrow stimulation and scaffold implantation in rabbits

Polymeric scaffolds

Macroporous scaffolds were copolymer networks made of EA and HEA [P(EA-co-HEA)]. Monomers were EA (99% pure; Sigma-Aldrich, St Louis, MO, USA) and HEA (96%, Sigma-Aldrich). Triethyleneglycoldimethacrylate (TEGDMA, 98%; Sigma-Aldrich) was used as cross-linking agent and benzoine (Scharlau, Sentmenat, Spain) as ultraviolet (UV) photosensitive initiator.

Scaffolds were prepared using a template made of PMMA microspheres with 90 µm average diameter (Colacryl DP 300, Lucite International, Southhampton, UK) that were sintered under pressure at temperatures above its glass transition temperature, according to the procedure previously described (36). The microspheres were introduced into a mold consisting of 2 glass plates separated by a rubber ring. The mold was placed into a hot press at 170°C to allow PMMA microspheres to soften, and then compressed to form a template of approximately 2 mm thickness. The interconnection points between porogen microspheres create the pore throats in the scaffold (pores of 90 µm diameter, as the template used). Thus, pore interconnectivity is controlled by the pressure applied when producing the template.

Templates were immersed in the monomeric solutions, including a 5% weight TEGDMA as cross-linker and 1% weight of benzoine as photoinitiator, and polymerized under UV light at room temperature. After polymerization, the template was dissolved with acetone for approximately 48 h in a Soxhlet extractor. Then, scaffolds were immersed in a large amount of acetone and the solvent was slowly changed to water in order to avoid the collapse of the scaffolds. Replicas of the scaffolds were cut in slices of approximately 3 mm in diameter and 1 mm in thickness. The scaffolds were then dried in vacuo for 24 h at room temperature followed by 24 h at 50°C, and they were finally sterilized with gamma radiation (25 kGy) before being used. Copolymers were prepared to contain 90% EA and 10% HEA by weight. These scaffolds are slightly hydrophilic and thus bulk polymers are able to absorb 2.3% and 3.3% of water measured on dry basis when immersed in liquid water until equilibrium (27).

Scaffold porosity was determined from the differences between the weights of dry samples and samples whose pores were filled with water. In order to ensure that water penetrated into all the interconnected pores, dry samples were introduced into a glass vessel at a high vacuum, and water was injected into it (37). The volume fraction of pores in the scaffolds was 0.75 ± 0.03 in all samples.

Scaffold morphology was examined by scanning electron microscopy (SEM) (Fig. 1). Dry scaffold samples were fractured in liquid nitrogen, mounted on copper stubs, and gold sputtered. The samples were observed with a scanning microscope (JSM5410; Jeol, Tokyo, Japan) under an acceleration tension of 15 kV.

Mechanical properties of the biostable porous scaffolds

Compressive strength measurement was carried out in a Thermo-Mechanical Analyzer (TMA) (EXSTAR 6000; Seiko Instruments, Chiba, Japan) in control position mode. Initially, a strain of 2% was applied for 15 min in order to adapt the sample surfaces to the probe. Afterwards, 4 successive programs of compression to 15% strain and decompression to initial 2% strain with a rate of 20 µm/min were performed at room temperature. The samples were tested in both dry conditions and immersed in water. The apparent Young modulus was calculated from the slope of the stress-strain curves in the linear region. The results are expressed as the average value of 5 measurements with its corresponding standard deviation (mean ± SD).

Animals

Two-month-old male New Zealand rabbits, weighing 1.5 kg to 2.0 kg were obtained from Granjas San Bernardo S.L. (Tulebra, Spain) and kept under conventional housing conditions. Quarantine lasted 7 days. Animals were housed with appropriate bedding and provided free access to drinking water and food. Rabbits were kept in standard single cages under controlled temperature and light conditions.

Spanish guidelines for the care and use of laboratory animals were observed. The study protocol was approved by the Ethics Committee of the University of Valencia according to law 86/609/EEC and 214/1997 and decree 164/1998 of Generalitat Valenciana Government.

Scaffold implant

Rabbits were preanesthetized by subcutaneous injection of 15 mg/kg ketamine (Ketolar; Pfizer, Madrid, Spain) and intramuscular injection of 0.1 mg/kg Medetomidine (Domtor; Pfizer), and prepared before surgery (washed, shaved, etc.). Then, general anesthesia was induced by 4% isofluorane using a specially designed mask and maintained by administration of 1.5% isofluorane with O₂ (2 l/min). The surgical site was sterilized using iodine solution and rabbit non-sterile parts were covered with sterile drapes. Surgeons wore sterile coats and gloves, and all instruments were sterilized and kept sterile during surgery. Non-preseeded scaffolds were moistened with phosphate buffered saline (PBS), and vacuum was applied to assure liquid penetration into the pores before implanting.

An arthrotomy at the knee joint was performed through a medial longitudinal parapatellar incision. The medial capsule was incised and the patella laterally dislocated. A 3 mm steel trephine was used to create a 3-mm diameter and about 1-mm depth defect in the central articular surface of the femoral trochlear groove, which resulted in the removal of articular cartilage along with subchondral bone injury. The defect was cleaned and rinsed with sterile saline, and scaffolds were laid into the defect, aligned to surrounding articular surface. Blood was allowed to flow inside scaffolds, which were held in place by repositioning the patella within the femoral trochlear groove. Control animals were subjected to the same operation but no scaffold was implanted in the cartilage defect. Arthrotomy and skin were sutured with continuous stitches of 4/0 coated Vicryl (Johnson-Johnson, New Brunswick, NJ, USA ). Macroscopic pictures were taken throughout the surgical procedure of all animals, using a Leica DC150 camera (Wetzlar, Germany). After removal of the conformed anesthesia mask, rabbits were returned to their cages and allowed free activity in the cage. Postoperative analgesia consisted of intramuscular injection of 3 mg/kg Dexketoprofen (Enantyum; Menarini, Florence, Italy) on the surgery day and the same dose every 24 h for 3 days. At the end of surgery, 3 mg/kg intramuscular injection of gentamicine (Genta-Gobens; Laboratorios Normon, Madrid, Spain) was administered as antibiotic prophylaxis.

Rabbit chondrocyte harvesting and culture

Articular cartilage was obtained from knee joints of donor rabbits after their sacrifice with a lethal intravenous injection of anesthetic overdose in the auricular vein (500 mg/iv sodium thiopental; Tiobarbital, B. Braun Medical, Barcelona, Spain). Chondrocyte isolation and culture was carried out as previously published (23).

Briefly, cartilage was dissected from subchondral bone, finely diced and washed with Dulbecco’s modified Eagle’s medium (DMEM; Life Technologies, Thermo Fisher Scientific, Fife, WA, USA) supplemented with 100 U penicillin, 100 μg streptomycin (Biological Industries, Kibbutz Beit Haemek, Israel) and 0.4% fungizone (Gibco-Thermo Fisher Scientific). In order to isolate chondrocyte, diced cartilage was successively digested using supplemented DMEM and different enzymes. First, cartilage was incubated for 30 min with 0.5 mg/ml hyaluronidase (Sigma-Aldrich, St Louis, MO, USA) in a shaking water bath at 37°C. Then, hyaluronidase was removed and 1 mg/ml pronase (Merck; VWR International SL, Barcelona, Spain) was added. After 60 min incubation in a shaking water bath at 37°C, cartilage pieces were washed with supplemented DMEM. Medium was removed and digestion continued by the addition of 0.5 mg/ml of collagenase-IA (Sigma-Aldrich), incubated overnight in a shaking water bath at 37°C. The resulting cell suspension was filtered through a 70 µm pore nylon filter (BD Biosciences, San Jose, CA, USA) to remove tissue debris. Cells were centrifuged and washed with DMEM supplemented with 10% fetal bovine serum (FBS; Invitrogen, Thermo Fisher Scientific). Finally, isolated cells were immediately used for chondrocyte culture.

Isolated cells were plated in culture flasks (T75; Becton-Dickinson, East Rutherford, NJ, USA) at high density in culture medium (DMEM supplemented with 10% FBS and 50 µg/ml ascorbic acid; Sigma-Aldrich) at 37°C in a 5% CO2 humidified atmosphere, as previously published (23). Medium was changed every 2 to 3 days. After 7 to 14 days, adherent cells were harvested by incubation with trypsin-EDTA (Biological Industries), resuspended with a minimum volume of culture medium and counted. Synthetic scaffolds were placed on a 24-well polystyrene culture plate (Nunc A/S) and moistened with Hanks’ Balanced Salt Solution (Sigma-Aldrich). After removing excess of Hanks’ solution, cell suspension (10⁶ viable cells in 20 µL medium) was injected in the center of some of the scaffolds to allow cell infiltration into the porous structure. After 1 h incubation, scaffolds were changed to a new well and culture medium was gently added to ensure that the material was covered. After 3 days in culture, medium was replace by DMEM containing 1% insulin-transferrin-sodium selenite media supplement (BD Biosciences) and 50 µg/ml ascorbic acid, and preseeded scaffolds were cultured during 3 more days before implanting to animals (3-month preseeded group).

Animal sacrifice and tissue retrieval

Animals with non-preseeded scaffolds were randomly distributed into 5 groups, and were sacrificed at 1 and 2 weeks, and at 1, 3, and 12 months after implantation. Since superficial cartilage was repaired 3 months after implantation in most animals, this time point was chosen to compare with an extra group consisting of animals in which preseeded scaffolds were implanted as well as with control animals, both of which groups were also sacrificed 3 months after surgery. The number of animals per group is detailed in Table I.

At the corresponding time after implantation, the rabbits were sacrificed as described above. Knee articular cartilage was observed and macroscopic pictures were taken, as mentioned above. Special care was taken in order to keep the repaired defect at the center of the sample in order to asses cartilage repair by histological techniques.

Histological studies

Morphology was studied following standard histological procedures. Briefly, rabbit articulation specimens were rinsed with PBS and fixed with 4% formaldehyde at room temperature for 5 days. Then, samples were rinsed with PBS and immersed in Osteosoft decalcifier solution (Merck, Whitehouse Station, NJ, USA) during 5 weeks at room temperature. Specimens were cut through the middle of the scaffold (where the scaffold diameter measured approximately 3 mm), and each half was separately embedded in paraffin. Five-μm thick sections were obtained and stained with hematoxylin-eosin and Masson’s trichrome. The ability of chondrocytes to synthesize glycosaminoglycan within the porous scaffold was monitored by Alcian blue staining (pH 2.5), counterstained using Harris hematoxylin. Stained sections were analyzed under an optical microscope (DM 4000B; Leica), and pictures were taken using a camera (DFC 420; Leica).

Immunohistochemistry

Standard immunohistochemistry techniques were performed to detect collagen type II, osteocalcin, and Ki-67 expression. Monoclonal mouse anti-collagen II antibody (Calbiochem; Merck, Darmstadt, Germany) at 1:200 dilution, incubated at 4°C overnight, was used to study the synthesis of collagen type II. Rabbit cartilage areas surrounding the scaffold were used as positive control. Osteocalcin detection was performed using a monoclonal mouse anti-­osteocalcin antibody (R&D Systems, Minneapolis, MN, USA) at 1:100 dilution, incubated at 4°C overnight. Rabbit subchondral bone was used as positive control. Monoclonal mouse anti Ki-67 antibody (MIB-1, DakoCytomation; Dako, Glostrup, Denmark) at 1:50 dilution, incubated at room temperature during 60 min, was employed to detect proliferating cells. Section of human neuroblastoma was used as a positive control. As a negative control for each specific staining, the preimmune serum was substituted for the primary antibody.

Sections were deparaffined and rehydrated through graded ethanol, rinsed in distilled water and treated with 0.3% H2O2 and 10% normal horse serum to block endogenous peroxidase and nonspecific binding, respectively. Antigen retrieval for collagen type II and Ki-67 was performed by pressure cooker boiling for 3 min in 10 mmol/l of citrate buffer (pH 6.0). For osteocalcin detection, the slides were permeabilized using 0.1% Triton X-100 for 5 min and antigen retrieval was performed using 0.5% trypsin for 30 min at 37°C in a humidified chamber. Envision amplification system Dako (CytomationEnvision+System labeled polymer-HRP anti-mouse) was used, followed by revelation with 3,3’-diaminobenzidine (Dako) as chromogen according to the manufacturer’s instructions, which originated a brown staining in immunoreactive structures. Sections were finally counterstained with Mayer's haematoxylin (Sigma-Aldrich).

Mechanical properties of the biostable porous scaffolds

An example of the stress-strain curve for the porous scaffolds is shown in Figure 2. The first stress-strain curve is different to the other curves that can be attributed to the Mullins effect, also called softening by deformation (38, 39). After the first cycle, the stress-strain curves are practically identical, and a small permanent deformation can be observed in the curves, which could be attributed to plastic deformation of the trabeculae and microfractures of the material with the collapse of some pores.

The apparent Young’s modulus was obtained from the slope of the stress-strain curve in the linear region of the compression curve; it was 0.57 ± 0.10 MPa for the dry samples and 0.34 ± 0.07 MPa for the immersed samples. This result is very similar to the reported values of 0.41 ± 0.12 MPa for the Young’s modulus in rabbits (30) and 0.58 ± 0.17 MPa in humans (40).

When a macroporous scaffold is implanted in a tissue, its porous structure is first filled with physiological fluid and then gradually with newly formed tissue. The role of water filling the pores can be an important factor in the mechanical behavior in compression. In this case, due to the high porosity, the water can flow outwards from the samples, and then, the main contribution of the water is a decrease in the ­compression modulus due to the plasticization effect of the water absorbed in the polymer.

Morphological study

The evolution of the implanted scaffolds was studied by histological techniques throughout a 12-month period, and they were examined at 1 and 2 weeks, and at 1, 3, and 12 months after implantation. In general, scaffolds were progressively shifted from the articular surface towards subchondral bone (Fig. 3, macroscopic view) while they were invaded by cells that formed neotissue, which filled a large fraction of the pore volume. In parallel to this process, native articular cartilage proliferated from the edge of the excavated area and formed a neocartilage that regenerated the condylar surface defect (Fig. 4, microscopic view).

Macroscopic analysis

Macroscopic observation of the samples showed that at early stages after implantation (1 week), scaffolds remained aligned with native articular cartilage, contacting with subchondral bone, and no signs of integration with surrounding tissue were observed; on the contrary, scaffolds were delimited by a sharp bloody line (Fig. 3a), which gradually disappeared thereafter. By 1 month after implantation, native articular cartilage seemed to have proliferated, surrounding the perforated cavity and causing the scaffold to sink slightly. Three months after implantation scaffolds were easily observed; they showed continuity with the surrounding tissues and were no longer aligned with the native articular surface, since they were covered by a layer of cartilage on the surface and also by bone tissue between cartilage and the scaffold in 6 of the 8 samples (Fig. 3b). When subchondral bone covered scaffolds, a cartilage-like band seemed to link scaffolds with superficial articular ­cartilage in 3 of the 8 samples. Twelve months after implantation, scaffolds were more difficult to observe since they were mimicking their surroundings and thus exhibited good integration with subchondral bone tissue (Fig. 3c). Figure 3d shows the articular surface 3 months after scaffold implantation. The appearance of the implant site (which can be detected by a slightly lighter color than the surroundings) is very good with a smooth surface, since the scaffold is covered by a regenerated tissue layer, as described above.

Microscopic analysis

Microscopic study revealed that scaffolds showed a progressive shift from articular surface at the earliest stages (Fig. 4a), toward subchondral bone at 3 to 12 months after implantation (Fig. 4b, c), as macroscopically observed.

The polymeric material did not stain in these samples included in paraffin, and therefore it was observed as white spaces. It remained unaltered throughout the first 3 months after implantation without any sign of degradation, although a slight thinning and deformation of the scaffolds was observed 12 months after implantation (Fig. 4c).

Articular surface

At the earliest stages after implantation (1-4 weeks), scaffolds appeared at the surface of the defect. By the second week after implantation, a thin layer of amorphous tissue was already observed in most samples covering their surface (Figs. 4a and 5a), which probably arrived from synovial cells. One month after implantation, a well-delineated layer of cell-enriched tissue was observed covering all scaffolds (Fig. 5b), which remained aligned with condylar surface.

Meanwhile, the superficial host cartilage began a repair process by activating the chondrocytes located near the edge of the excavated tissue. As earlier as 1 week after implantation, lacunae in the native cartilage showed an increased size and proliferation of chondrocytes, with the appearance of more than 10 cells inside some lacunae (Figs. 6a and b). Along the time studied, the neocartilage matrix formed was growing centripetally from the edge of the excavated native cartilage toward the center, so that it seemed to be pushing the scaffold toward subchondral bone (Figs. 6c and d). Therefore, 3 months after implantation 6 of the 8 samples showed that the growing neocartilage had arrived at the center of the excavated area, contacting and merging with the opposite edge, thus covering the whole scaffold surface at this time (Figs. 7a-c) and partially in the other 2 of the 8 samples. Small areas of fibrous cartilage (Fig. 7b) were observed in 2 of the 8 samples at this time point. Twelve months after implantation, all samples studied showed excellent cartilage repair, with a well-organized layer of hyaline-like cartilage on the top of scaffolds (Fig. 7d), similar to non-treated cartilage.

Nesting cells

At the same time that neocartilage was growing on the surface, incipient tissue formation could be observed within scaffold pores from the earliest stages after implantation, which was likely the result of differentiation of mesenchymal cells arriving from subchondral bone marrow. Neotissue formation started 1 week after implantation in the pores located in the deepest zone of the scaffold (Fig. 8a), and spread out throughout the scaffold, arriving at the superficial zone by 1 month ­after surgery. The amount and type of neotissue formed varied over time and according to the location within the scaffolds. At the beginning, few fibroblast-like, morphologically ­undifferentiated cells were observed within pores, surrounded by scarce amounts of homogeneous non-fibrous extracellular matrix (Fig. 8a). By 2 weeks after implantation, an increasing number of undifferentiated cells was observed inside the pores along with small capillaries, which were occasionally observed emerging from subchondral bone (Fig. 8b). One month after implantation, the newly formed tissue showed signs of morphological differentiation, with a round-cell shape, and a more abundant extracellular matrix that occupied more extensive pore areas than those previously observed (Figs. 8c and d). By the third month after implantation, scaffold pores were apparently filled mostly with cartilaginous tissue in the superficial and middle zones of the scaffold (Fig. 8e), as revealed by Alcian blue staining (not shown) and type II collagen immunostaining (Fig. 9a), whereas bone tissue seemed to be the predominant in the deeper zone, adjacent to the subchondral bone (Figs. 8f and 9b). Abundant proliferating cells were observed at this time, as could be checked by Ki-67 immunostaining (Fig. 9c). Twelve months after implantation, when scaffolds were located at great distance from the articular surface (Figs. 4c and 7d), bone was the predominant tissue observed infiltrating scaffolds.

Surroundings

Scaffold integration in the surrounding tissue started as soon as mesenchymal cells invaded scaffold pores (Figs. 8a and b). Three months after implantation, when scaffolds were fully included in the joint tissue and the articular surface was repaired, scaffolds were partly integrated into the surrounding tissue, with continuity between the neotissue grown inside the scaffold pores and the surrounding native tissue, in about 50% of their outlines (Fig. 8e). No necrosis was found, but occasionally, areas of fibrous, inflammatory or reactive tissues, containing leucocytes or phagocytic cells, were also observed surrounding smaller areas (about 20% of their outlines) in 5 of the 8 scaffolds (data not shown). Twelve months after implantation, scaffolds were surrounded by spongy bone with microscopic characteristics of lamellar bone tissue (Fig. 7d).

Preseeded scaffolds

The repair tissue obtained 3 months after implantation was compared with that obtained when scaffolds were preseeded with allogenic chondrocytes. In this case, scaffold behavior resembled non-preseeded ones, since they were also located at certain distance from the articular surface, the superficial cartilage was similarly regenerated, and the integration with its surroundings looked like the non-preseeded samples. Subchondral bone covered the entire scaffold upper surface in 1 of the 5 samples, and only 25% to 50% of that surface in the other 4 samples (Fig. 10). Moreover, the quality of the tissue grown inside the scaffold pores was similar, although preseeded scaffolds seemed to have a higher proportion of neotissue filling their pores (Fig. 10).

Control group

The three-month group was chosen to compare cartilage repair to controls, which underwent the same surgerical procedure, but without scaffold implantation in the chondral defect. Histological study of the tissue formed in the control animals showed the typical aspect of the regenerated cartilage after microfracture or analogous procedures to injure subchondral bone and allow bleeding into the cartilage defect (41). The articular surface looked rough and irregular macroscopically, whereas disorganized cells were observed under microscope filling the excavated cavity, distributed in an extracellular matrix with a fibrous cartilage aspect (Fig. 11). However, no appearance of growing edges was observed at the host cartilage, as it was observed in the scaffold-implanted samples.

Articular cartilage repair

Histologically, articular cartilage is a hyaline cartilage divided into 4 zones, distinguished by the shape of the chondrocytes and the composition of extracellular matrix including the arrangement of collagen fibers (3, 4, 43), as shown in Figure 12. The superficial zone (SZ) is characterized by small-flattened chondrocytes, low proteoglycan content, and densely packed, horizontally arranged collagen fibrils of uniform diameter. In the middle or transitional zone (MZ), chondrocytes acquire a more rounded profile, proteoglycan content increases, and thick collagen fibers decussate to provide an oblique transitional network between the superficial and deeper zones. The radial zone (RZ) is characterized by spheroid chondrocytes arranged in columns, high proteoglycan content between radially oriented thick collagen bundles. A microscopically distinct line – the tidemark (T) – separates the lower radial zone from the underlying calcified cartilage zone (CZ), which lies above the subchondral bone.

Adult articular cartilage has a very low self-repair ability (3), although it has an active proliferation during prenatal development, which decreases precipitously after birth in New Zealand white rabbits, and undergoes a process of fundamental reorganization during the first 3 months of the postnatal period. Thus, a rapidly proliferating cell population is present in articular cartilage of 2-month-old rabbits (5), located in the middle and upper radial zones (transit-amplifying cells), enlarging epiphysis not only in a longitudinal direction, but also radially and laterally (4). Therefore, proliferation and reorganization of neotissue should occur in the animals used in the present study, either in the presence or the absence of scaffold. In fact, control animals (in which no scaffolds were implanted) showed the ability to produce newly formed tissue but with the structure of fibrous cartilage, in good agreement with the literature. This newly formed tissue can be distinguished under the microscope since hyaline cartilage exhibit a homogeneous extracellular matrix (Fig. 12), whereas collagen fibers are easily distinguished in the fibrous cartilage (Fig. 11b).

The repair of articular cartilage we observed agrees with previous studies where hyaline cartilage was regenerated on the top of the scaffolds, using a similar protocol but with different biomaterials (30). This suggests that the fast growth of this tissue layer had embedded the scaffold toward the subchondral bone. Indeed, in the early stages after implantation we observed a large number of chondrocytes in many lacunae of the surrounding native cartilage that later on synthesized cartilaginous matrix, initiating the overall growth of the articular cartilage. However, Shapiro et al (41) demonstrated that residual adjacent articular cartilage did not participate in the repopulation of the defect, but unlike our study, they did not implant any scaffold in the defect.

As stated above, although excellent cartilage repair was observed 1 year after surgery, this must be partly attributed to the age of the animals at the time of surgery (2 months), when cartilage reorganization is still very active (4). However, a young age was not sufficient for cartilage regeneration, since the absence of scaffold highly influenced its functionality and initiated cartilage growth but with a fibrous appearance in the control animals.

Mechanical stress support

The function of articular cartilage is intimately linked to biomechanics, since it performs the function of load support and lubrication with minimal wear and little or no damage in animals and humans (1). Thus, biomaterial scaffolds should, on the one hand, provide the chondrogenic cells with a microenvironment where they survive, multiply, and produce extracellular matrix to constitute regenerated cartilage (1, 44). On the other hand, scaffolds should have mechanical properties matching those of native cartilage (15, 16). Accordingly, we used P(EA-co-HEA) copolymer with 5% weight TEGDMA as cross-linker in order to obtain a Young's modulus similar to that of native cartilage. Increasing the amount of TEGMA produces a rise in the Young's modulus, but also brittle material, whereas a smaller amount of TEGMA produces a weaker material.

The scaffold material used in the present study is biostable and thus it remains almost unaltered 12 months after implantation. Therefore, scaffolds are biomechanically active throughout the regenerative process, and the cells within scaffold pores are subjected to compression stresses similar to those suffered by the chondrocytes of healthy tissue (44). This is important in the differentiation process of chondrocytes in order to acquire the phenotype of hyaline cartilage cells (45).

In previous studies we used scaffolds made of polycaprolactone, empty or preseeded with cultured chondrocytes, and we observed that defects were filled with cartilaginous tissue 3 months after implantation in a way that is similar to the present study (30). This suggests that the mechanical role of the scaffold is to provide a substrate that helps the transfer of the compressive stresses to the newly formed tissue.

However, when subchondral bone was injured but no scaffold was implanted (control animals), the repaired articular cartilage 3 months after surgery was not hyaline cartilage but fibrous, in good agreement with previous studies (30, 41). It is also in accord with a long-term study by Hunziker and Rosenberg, who observed the appearance of fibrous connective tissue 1 year after an articular cartilage defect that was filled only with mitogenic factors or with cells (46). This result suggests the importance of the mechanical transduction performed by the scaffold in fostering an adequate biomechanical environment for the repaired tissue.

The cells located on the surface of the scaffold, aligned with the articular surface, are subjected to compressive loading that directs the extracellular matrix organization and probably plays an important role in chondrogenic differentiation through mechanotransductive pathways. This biomechanical environment is maintained while the cartilage surface layer grows, pushing the scaffold into subchondral bone. The presence of the scaffold is also crucial to initiate bone remodeling at the places where it contacts and transmits compression loading to subchondral bone (47). Therefore, although the ability of cartilage repair is not entirely lost at this age, our results suggest that the scaffold plays a key role in stress transmission to the cells, which is a vital factor in cell phenotype maintenance and tissue organization.

Neotissue growth within scaffolds

Scaffold pores were invaded by mesenchymal cells that proliferated and morphologically differentiated to cartilaginous or bone tissues 3 months after implantation. This result is in agreement with other studies (41, 48), as well as with similar studies that used scaffolds made of different materials in the presence or absence of preseeded chondrocytes (30, 32).

The origin of cells that colonize the scaffolds can vary. The main cellular source is the outflowing bone marrow blood that contains pluripotent stem cells, which are established by the clot formation in the defect. This provides an enriched environment for tissue regeneration, making these cells a potential source for cartilage repair (2, 8, 11).

The arrival of mesenchymal cells to the injured zone can lead to cartilage regeneration if a number of circumstances occur (13), such as: initial cell recruitment to the injury site; cell adhesion to a local matrix followed by activation and extensive cell proliferation; cell switch to chondrogenic matrix production, whose composition changes during the differentiation of mesenchymal cells into chondrocytes (49); integration with neighboring tissue; and finally regeneration of a tidemark, adaptation to biomechanical loading and building up a balance tissue homeostasis.

Synovium is also a potential source of mesenchymal cells for clinical applications. It has been shown that synovial-derived mesenchymal cells have a great ability for chondrogenesis, suggesting that they may be a more attractive source for cartilage repair than bone marrow (46, 50, 51). This also suggests that they may have been partly responsible for the repair observed in the present study.

Cartilage integration in the surrounding tissue is one of the major problems that limit the efficacy of cartilage repair (46). The porous architecture of the scaffolds used in the present study facilitates its integration with adjacent cartilage and subchondral bone tissues, besides providing a favorable environment for neotissue proliferation, migration and differentiation state maintenance. In the present study, we observed that scaffolds were well anchored to the adjacent host tissue 3 months after implantation in most cases, which can be attributed to the ingrowth and remodeling of surrounding tissue into the scaffold. Different results have been shown in similar studies, for instance Schaefer and collaborators reported that bone tissue also showed good integration whereas cartilage did not (52), while Wang et al reported a better surrounding cartilage infusion (32). These results are probably due to the differences of the material used and the age of the rabbits at the time of surgery.

We compared tissue repair 3 months after surgery to an additional group where scaffolds were preseeded with allogenic chondrocytes before implantation. Preseeded scaffolds showed no qualitative differences with respect to the non-preseeded ones at the same time after implantation, although the amount of tissue grown within the scaffolds seemed greater when the scaffolds were preseeded with cultured chondrocytes. This result agrees with other studies where similar protocols were followed (30).

Autologous chondrocyte transplantation is a suitable technique for improving the rate of repair of large articular cartilage defects (48). However, this technique requires obtaining cells from a donor and sufficient cell expansion, which can be accompanied by chondrocyte dedifferentiation – with the production of a non-hyaline, cartilaginous, mechanically inferior extracellular cartilage matrix (48). For this reason, special care should be taken in maintaining chondrocyte (re)differentiation in order to obtain chondral matrix. Thus, since superficial cartilage repair, neotissue growth and integration with the surroundings showed only slight differences between non-preseeded and preseeded scaffolds at the age studied, we suggest that the use of the preseeded materials is outbalanced by the far more complicated and risky procedure.